$ npm install -g @openai/codex@0.149.1ChatGPT & Codex changelog
Latest updates to ChatGPT and Codex
August 2026
-
Codex CLI 0.149.1
-
Codex and ChatGPT updates
New features
-
Apple Messages: Use the Apple Messages plugin to read and search Messages chats on your Mac and prepare or send messages. It’s available on all plans in the ChatGPT desktop app for macOS. You can use the plugin in ChatGPT Work and Codex. By default, ChatGPT sends messages only after you approve the message and its recipients. See the plugin guide for persistent-approval risks, revocation steps, and the known issue with tasks that disable approval prompts.
-
Site co-editing: Where Site collaboration is available, owners can invite active members of the same workspace as editors. Editors can read the Site’s live database data, update the Site, save versions, and publish changes after the owner publishes the Site for the first time. Owners retain control of the audience, settings, analytics, ownership, version restoration, and editor access.
-
Editable Site URLs: Where URL editing is available, owners can change an existing Site’s ChatGPT-hosted address without creating another deployment. The previous address redirects to the new URL. Custom domains are a separate, existing feature and aren’t changed by this setting.
-
Computer History in Europe: Computer History is now available in the EEA, Switzerland, and the United Kingdom for ChatGPT Pro, Business, and Enterprise users in the ChatGPT desktop app on macOS. It’s off by default and requires Memories. Business and Enterprise administrators must enable access before workspace members can choose to turn it on.
-
Shared thread snapshots: On all Codex plans, share a read-only snapshot of a local Codex thread from the ChatGPT desktop app for macOS. The snapshot doesn’t update when the original thread changes. Personal-account links can be opened by anyone with the link; workspace-account links are limited to members of the originating workspace. Codex redacts known secret patterns, but review the shared content because sensitive content may remain. View or revoke links in ChatGPT data controls, under Shared links.
-
Unified pinned threads: Keep the same pinned chats across the ChatGPT desktop app and iOS.
-
-
Codex CLI 0.149.0
$ npm install -g @openai/codex@0.149.0View details
New Features
- Added an interactive
codex agentsdashboard for searching, starting, opening, renaming, and stopping tasks, with configurable shortcuts. (#39094, #39112, #39114, #39142) - Added
/cd,/pwd, and/cwdcommands for managing the working directory in TUI sessions. (#38894) - Added
codex queuefor sending messages to existing local or remote sessions. (#39092) - Expanded Vim editing with character replacement and more change motions such as
cw,c$, andcc. (#39661) codex doctornow diagnoses endpoint protection, network/proxy failures, desktop app state, and update connectivity. (#38827, #38918, #39060, #39074)- SDK users can now pass exact CLI config overrides and select
maxorultrareasoning effort. (#38817, #39662)
Bug Fixes
- Queued messages now wake idle sessions reliably, resolve duplicate session names more usefully, and preserve pasted or deferred command semantics. (#39034, #39385, #39604)
- Resumed and forked threads now restore their active permission profile instead of silently falling back to current defaults. (#39153)
- Fixed duplicate sub-agent activity and tightened TUI routing for sub-agent notifications and approvals. (#39049, #39088)
- Realtime WebRTC sideband connections now reconnect after unexpected transport loss without dropping pending output. (#39257)
- Inline TUI history now remains available in Windows Terminal scrollback. (#39619)
- Inactive TUI thread replay buffers are now bounded to prevent excessive retained streamed output. (#39081)
Documentation
- Clarified that external contributions should go through issues and design discussion rather than pull requests. (#39089)
- Documented DNS exfiltration risks and trust limitations for secure devcontainers. (#39283)
Changelog
Full Changelog: rust-v0.148.0...rust-v0.149.0
- #38817 Add raw config overrides to the TypeScript SDK @copyberry
- #38819 Support metadata staging for reserved thread IDs @copyberry
- #38822 Avoid cloning TUI history span content @copyberry
- #38823 Avoid allocating per character when decorating hyperlinks @copyberry
- #38827 Add endpoint protection checks to
codex doctor@copyberry - #38830 Isolate external editor buffers from sandbox-writable paths @copyberry
- #38837 Share editor keymaps across TUI composer components @copyberry
- #38840 Identify Mac mini hosts in remote control handshakes @copyberry
- #38893 Restore thread timestamp maxima independently @copyberry
- #38894 Add working-directory commands to the TUI @copyberry
- #38899 Move requirements policy ownership to execpolicy @copyberry
- #38902 Honor per-environment shell variable policies @copyberry
- #38907 Edit queued messages with Vim history-up @copyberry
- #38913 Stop rendering columns after filling their area @copyberry
- #38916 Honor legacy
:project_rootspermission entries @copyberry - #38918 Improve
codex doctornetwork diagnostics @copyberry - #38919 Reject obsolete app-server permission profile fields @copyberry
- #38921 Compact successful command activity in the TUI @copyberry
- #38940 Add experimental app-server project APIs @copyberry
- #38941 Prevent Noise auth tokens from reaching child processes @copyberry
- #38942 Enforce environment-specific command policies @copyberry
- #38947 Stop loading legacy managed config on Windows @copyberry
- #38978 Add a configurable skill catalog token budget @copyberry
- #38980 Bound Guardian v2 parent compaction context @copyberry
- #38987 Include images in Guardian v2 transcripts @copyberry
- #38990 Source Guardian v2 defaults from the model catalog @copyberry
- #38993 Evaluate recent and character-routed skill selection @copyberry
- #38994 Remove the workspace settings gate for apps and plugins @copyberry
- #38997 Retry transient Guardian sampling failures @copyberry
- #39001 Ignore stale Guardian tool risk scores @copyberry
- #39005 Disable Guardian V2 for managed automatic reviewers @copyberry
- #39006 Emit API-key turn cost telemetry from app-server @copyberry
- #39007 Wait for Guardian score progress in approval tests @copyberry
- #39008 Add task context to shadow skill selection @copyberry
- #39017 Move MCP policy types into
codex-protocol@copyberry - #39019 Isolate MCP resource headers during OAuth requests @copyberry
- #39020 Restrict Windows IDE pipe client impersonation @copyberry
- #39022 Hang soft-break whitespace in the TUI composer @copyberry
- #39023 Reduce retained MCP result data in TUI history @copyberry
- #39025 Speed up the large diff highlighting test @copyberry
- #39026 Use simulated time in the long typing test @copyberry
- #39028 Cache the test pet spritesheet encoding @copyberry
- #39030 Speed up the in-app resume cwd test @copyberry
- #39032 Encapsulate TUI shimmer under the motion module @copyberry
- #39033 Bound legacy resume preview scans @copyberry
- #39034 Dispatch queued messages written by other processes @copyberry
- #39035 Add app-server coverage for Guardian V2 approval routing @copyberry
- #39036 Allow config reads to join active app-server read batches @copyberry
- #39038 Restore Guardian risk scores across thread lifecycles @copyberry
- #39040 Preserve foreign paths in managed network approvals @copyberry
- #39043 Enforce managed authentication backend settings @copyberry
- #39045 Add managed gates for in-app chat and dictation @copyberry
- #39046 Restrict MCP HTTP redirects to the configured origin @copyberry
- #39047 Skip plugin hook loading when hooks are disabled @copyberry
- #39049 Avoid rendering sub-agent activity twice in the TUI @copyberry
- #39050 Prepare the telemetry shutdown worker during initialization @copyberry
- #39051 Use installed callable apps for TUI mentions @copyberry
- #39055 Add network policy metadata to environment configuration @copyberry
- #39056 Verify the pinned zsh manifest in release builds @copyberry
- #39057 Rate-limit TUI frames from their actual emission time @copyberry
- #39058 Tag Codex Apps protocol discovery metrics @copyberry
- #39060 Add desktop app diagnostics to
codex doctor@copyberry - #39061 Avoid rerendering streamed code fences @copyberry
- #39063 Render only visible rows in the transcript pager @copyberry
- #39064 Restrict queued-message editing to its dedicated binding @copyberry
- #39065 Limit terminal hyperlink layout to the visible viewport @copyberry
- #39067 Add desktop security enforcement diagnostics @copyberry
- #39068 Remove skill model delegation support @copyberry
- #39072 Persist generated images through turn executors @copyberry
- #39073 Propagate caller metadata to rendezvous connections @copyberry
- #39074 Add desktop update diagnostics to
codex doctor@copyberry - #39075 Avoid redundant terminal row clears @copyberry
- #39077 Build filesystem JSON params only for remote TUI sessions @copyberry
- #39078 Preserve tracing context for environment resolution @copyberry
- #39079 Apply user MCP policy to selected executor plugins @copyberry
- #39081 Bound TUI thread replay buffers by delta size @copyberry
- #39082 Prompt for project trust in remote TUI workspaces @copyberry
- #39083 Harden Windows sandbox provisioning against reparse points @copyberry
- #39084 Preserve filesystem permission path conventions @copyberry
- #39087 Read plugin authentication state from AuthManager @copyberry
- #39088 Harden TUI subagent navigation @copyberry
- #39089 Clarify the external contribution policy @copyberry
- #39092 Add a command to queue messages for existing sessions @copyberry
- #39094 Add an agents overview dashboard to the TUI @copyberry
- #39098 Trace exec-server requests from receipt through completion @copyberry
- #39100 Avoid redundant terminal size queries during history insertion @copyberry
- #39101 Update rmcp to 3.1.2 @copyberry
- #39102 Raise the GPT-5.6 maximum context window @copyberry
- #39103 Drop capabilities from Linux sandbox processes @copyberry
- #39112 Make the agents overview an interactive task dashboard @copyberry
- #39113 Surface interactive requests in realtime conversations @copyberry
- #39114 Add a dedicated
codex agentsdashboard command @copyberry - #39115 Remove the experimental thread config endpoint @copyberry
- #39117 Reject lossy legacy permission projections @copyberry
- #39122 Fail closed on deeply nested command wrappers @copyberry
- #39131 Validate identifiers in plugin creator workflows @copyberry
- #39141 Redact auth tokens from app-server response logs @copyberry
- #39142 Add configurable shortcuts for the agents dashboard @copyberry
- #39143 Hydrate recommended plugin metadata on selection @copyberry
- #39145 Persist active permission profiles in turn context @copyberry
- #39147 Centralize persisted resume settings lookup @copyberry
- #39152 Update PyPI publish action to v1.14.2 @copyberry
- #39153 Restore permission profiles when resuming threads @copyberry
- #39154 Box the TUI future to bound CLI stack usage @copyberry
- #39155 Prepare Python SDK 0.147.0 stable release @copyberry
- #39157 Notify clients when Guardian requires strict review @copyberry
- #39159 Require approval for commands with dynamic shell words @copyberry
- #39163 Refresh collaboration instructions when their content changes @copyberry
- #39165 Prevent marketplace identity spoofing @copyberry
- #39174 Skip empty user messages for automatic idle turns @copyberry
- #39176 Drop descendant progress updates after remote compaction @copyberry
- #39187 Scope MCP app resource reads to their originating call @copyberry
- #39192 Preserve MCP resource origins across compaction @copyberry
- #39200 Add a symlink-safe reader for sensitive files @copyberry
- #39205 Reject symbolic links in memory workspaces @copyberry
- #39213 Add a fail-closed Tree-sitter PowerShell lowerer @copyberry
- #39214 Prevent custom providers from inheriting ambient auth @copyberry
- #39220 Reconnect Guardian sampling WebSockets after auth changes @copyberry
- #39221 Skip redirected external-agent migration destinations @copyberry
- #39224 Add Guardian v2 approval review metrics @copyberry
- #39227 Include node_repl screenshots in Guardian v2 reviews @copyberry
- #39235 Decouple Noise relay streams from JSON-RPC processing @copyberry
- #39240 Deduplicate remote plugin bundle syncs with shared semaphores @copyberry
- #39241 Record Guardian v2 classification metrics @copyberry
- #39242 Add safe permission profile intersection @copyberry
- #39244 Scope MCP resource reads by connector @copyberry
- #39246 Give Guardian classifier connections distinct thread identities @copyberry
- #39249 Add exec-server forwarding mode @copyberry
- #39256 Deduplicate rollout moves when archiving threads @copyberry
- #39257 Reconnect WebRTC Realtime sideband transports @copyberry
- #39259 Simplify unified exec output snapshots @copyberry
- #39261 Stop TUI chats on misalignment policy violations @copyberry
- #39262 Prevent ConPTY DLL loading from the current directory @copyberry
- #39264 Improve Guardian v2 risk classification @copyberry
- #39266 Require fresh approval beneath denied permission paths @copyberry
- #39267 Inject Node REPL policy into Guardian review sessions @copyberry
- #39273 Preserve thread names during rollout migration @copyberry
- #39274 Add provider-owned authentication recovery @copyberry
- #39277 Declare experimental Amazon Bedrock setup APIs @copyberry
- #39278 Preserve owner-provided environment configuration @copyberry
- #39279 Propagate Windows sandbox ACL update failures @copyberry
- #39281 Move shell snapshot scripts into
codex-shell-command@copyberry - #39283 Document secure devcontainer DNS exfiltration risk @copyberry
- #39284 Report network disconnects during approval @copyberry
- #39285 Show file destinations in TUI change approvals @copyberry
- #39287 Report diagnostic upload failures @copyberry
- #39288 Register the async message feature flag @copyberry
- #39290 Add Windows sandbox diagnostics to
codex doctor@copyberry - #39293 Remove app-server's direct reqwest dependency @copyberry
- #39294 Increase SQLite log sink batching @copyberry
- #39296 Enable MCP tool hooks in Codex sessions @copyberry
- #39298 Allow overriding Codex package versions @copyberry
- #39299 Restrict agent roles to bounded configuration overrides @copyberry
- #39301 Prevent Node REPL auth tokens from reaching child processes @copyberry
- #39303 Record Guardian v2 classification token usage @copyberry
- #39304 Keep Guardian v2 risk scores in memory @copyberry
- #39306 Honor managed config during project discovery @copyberry
- #39307 Fail closed on Guardian V2 risk scoring errors @copyberry
- #39309 Attribute executor skill invocations to plugins @copyberry
- #39311 Bind unified exec approvals to shell executables @copyberry
- #39312 Add async delivery metadata to agent messages @copyberry
- #39314 Run hooks with the captured session environment @copyberry
- #39315 Evict guardian transcript entries in cacheable chunks @copyberry
- #39316 Support Edu Plus and Edu Pro account plans @copyberry
- #39319 Add the async user message tool @copyberry
- #39320 Expand OAuth metadata redirect test coverage @copyberry
- #39322 Enforce workspace restrictions for header authentication @copyberry
- #39325 Stop migrating Cursor sandbox settings @copyberry
- #39331 Route hook MCP calls through current connections @copyberry
- #39335 Enforce environment MCP policies @copyberry
- #39372 Scope TUI approval requests to their threads @copyberry
- #39385 Prefer the most recent session when queueing by name @copyberry
- #39402 Remove npm package staging from repo checks @copyberry
- #39404 Support FD mounts with older system Bubblewrap versions @copyberry
- #39410 Refresh expired AWS credentials for Bedrock @copyberry
- #39452 Remove the feature gate for async user messages @copyberry
- #39474 Consolidate Guardian extensions into
codex-guardian-v2@copyberry - #39480 Move shell snapshot tests into shell-command @copyberry
- #39493 Make head-tail buffer capacity const generic @copyberry
- #39494 Test panoramic Guardian transcript image resizing @copyberry
- #39496 Use default timeouts in cyber exec policy tests @copyberry
- #39497 Correct normalized dynamic tool coverage across response modes @copyberry
- #39501 Use a narrow fixture for the unified image resize test @copyberry
- #39505 Test text stringify errors in the code mode runtime @copyberry
- #39506 Test code mode notifications without a sync tool call @copyberry
- #39509 Test disabled enhanced Node REPL transcript images separately @copyberry
- #39510 Track built-in control tool calls in analytics @copyberry
- #39514 Use stored item types when materializing turn summaries @copyberry
- #39515 Use
mem::taketo drain unified exec output buffers @copyberry - #39520 Isolate automatic plugin Git operations @copyberry
- #39523 Persist thread section moves before the first turn @copyberry
- #39524 Stop treating Git commands as inherently safe @copyberry
- #39584 Add a just recipe for assembling Codex packages @copyberry
- #39585 Test plugin sync isolation from repository Git config @copyberry
- #39586 Isolate IPC in Bubblewrap sandboxes @copyberry
- #39588 Preserve unparsed shell wrappers in exec policy @copyberry
- #39590 Harden plugin manifest handling during installation @copyberry
- #39592 Prevent SQLx warnings from feeding back into SQLite logs @copyberry
- #39594 Raise the MCP tool name limit to 128 bytes @copyberry
- #39595 Keep marketplace upgrade state out of config @copyberry
- #39597 Separate thread settings from environment configuration @copyberry
- #39599 Protect macOS Seatbelt writable root anchors @copyberry
- #39601 Keep async user messages on the direct tool surface @copyberry
- #39602 Use in-process parsing for PowerShell command classification @copyberry
- #39604 Preserve queued TUI input semantics @copyberry
- #39605 Hide approved automatic review warnings in the TUI @copyberry
- #39606 Enable user namespaces in shared CI setup @copyberry
- #39607 Resolve model-provided shells by type @copyberry
- #39608 Harden skill installation against unsafe symlinks @copyberry
- #39609 Limit Bazel integration test threads on macOS @copyberry
- #39611 Harden MCP OAuth fallback credential writes @copyberry
- #39614 Prevent
apply_patchfrom widening write permissions @copyberry - #39615 Bind MCP OAuth refresh tokens to their issuer @copyberry
- #39616 Validate linked worktrees before inheriting project trust @copyberry
- #39618 Apply composer editing preferences to TUI text prompts @copyberry
- #39619 Preserve inline TUI scrollback in Windows Terminal @copyberry
- #39620 Stream executor capability and skill file reads @copyberry
- #39623 Prevent protected-path rename bypasses in macOS Seatbelt @copyberry
- #39625 Add cwd-relative turn diff paths @copyberry
- #39629 Preserve parent repository discovery through sandbox metadata mounts @copyberry
- #39630 Retire the untrusted approval policy @copyberry
- #39631 Skip sandboxed shell commands in Guardian v2 by default @copyberry
- #39632 Expose permission profile resolution in the core API @copyberry
- #39635 Show strict review warnings in the TUI @copyberry
- #39637 Treat
invalid_grantrefresh failures as permanent @copyberry - #39640 Prompt to unarchive sessions before resuming or forking @copyberry
- #39641 Sanitize developer context in full-history agent forks @copyberry
- #39645 Enforce managed residency for model providers @copyberry
- #39646 Exercise restricted-token sandboxing in cyber policy tests @copyberry
- #39649 Resolve bundled Windows helpers through bin junctions @copyberry
- #39653 Enforce filesystem permissions when loading AGENTS.md @copyberry
- #39655 Make core integration test permissions explicit @copyberry
- #39656 Advertise the Desktop app in graphical Linux sessions @copyberry
- #39657 Warn when launching the deprecated MCP server @copyberry
- #39658 Let Guardian V2 satisfy required model reviews @copyberry
- #39659 Harden unsandboxed patch filesystem access @copyberry
- #39661 Expand Vim change commands and add character replacement @copyberry
- #39662 Add max and ultra reasoning efforts to the SDKs @copyberry
- #39663 Restrict plugin migration to home scope @copyberry
- #39665 Add macOS Seatbelt filesystem integration tests @copyberry
- #39666 Improve no-follow filesystem behavior across platforms @copyberry
- Added an interactive
-
GitLab support in Codex cloud (Beta)
GitLab support is available in beta on all ChatGPT plans. Connect a GitLab project to Codex cloud, create an environment for it, start tasks from issues or merge requests with
@codex, and request one-off or automatic merge request reviews.The integration runs in Codex cloud. A managed workspace admin can disable the connector. GitLab-triggered activity requires permission to configure the applicable webhook. For GitLab Self-Managed or GitLab Dedicated, a workspace admin must configure the connection, and webhook activity requires GitLab 19.0 or later.
Codex cannot complete a review when GitLab omits a collapsed or oversize diff.
Learn how to use Codex with GitLab.
-
ChatGPT for iOS 1.2026.223
New features
- Added a setting to open ChatGPT directly in Codex Remote on launch.
- Added support for standard MCP forms and editable message approvals.
- Linked folders now open directly in the files sheet.
Improvements and bug fixes
- Improved the New Thread project picker to reflect the selected host’s current projects.
- Voice now works directly from existing task composers, connects more reliably, and continues task actions in the background.
- Improved diff review stability and performance, especially in large workspaces.
- Added a Retry action when task messages fail to load.
- Fixed large task responses failing to load.
- Fixed tasks disappearing or remaining unavailable after being idle, reconnecting, or returning to the task list.
- Improved host pairing reliability and prevented enrollment checks from freezing the app.
- Improved response annotations and preserved streamed activity when tasks complete.
- Canceling or editing a steering message now prevents delivery.
-
Codex CLI 0.148.0
$ npm install -g @openai/codex@0.148.0View details
New Features
- Export complete TUI conversations to Markdown with
/export, either to the clipboard or a new file. (#37358) - Fork sessions with
codex exec fork, and archive or restore sessions from the TUI resume picker. (#37367, #37369, #37371) - Draft prompts while the TUI initializes, with resume and fork progress shown during startup. (#38642, #38788)
- View estimated thread credits or cost in
/status, status lines, and terminal titles for eligible workspaces. (#38281, #38282) - Use Amazon Bedrock Runtime as a built-in provider with AWS profile, region, and GPT-5.6 routing support. (#38470)
- Hooks can now run commands asynchronously and invoke MCP tools. (#37533, #38705)
Bug Fixes
- Model switches and settings updates no longer leave stale instructions behind or change an active turn midstream. (#37260, #38785)
- Resumed sessions now restore their persisted working directory and approval policy, with more accurate transcript previews. (#37198, #37368, #38605)
- Turns reconnect through temporary provider outages, and MCP servers recover after OAuth reauthentication without restarting Codex. (#37337, #37485, #38418)
- TUI startup no longer lets buffered terminal input activate prompts accidentally, and onboarding appears when authentication is missing. (#38641, #38643, #38644)
- Composer and transcript rendering now handle CRLF pastes, wrapped whitespace, and long URLs correctly. (#37709, #38380, #38704)
- Sandbox restrictions now fail closed for denied or unreadable paths across Linux and Windows. (#37875, #38026, #38416, #38660)
Documentation
- The bundled skill-creator guide is more focused, and validation now rejects unfinished TODO placeholders. (#38384)
Changelog
Full Changelog: rust-v0.147.0...rust-v0.148.0
- #37190 Interrupt cyber model turns after one Guardian denial @copyberry
- #37191 Preserve legacy semantics during rollout migration @copyberry
- #37198 Prefer persisted cwd when reading local threads @copyberry
- #37199 Track thread archive analytics @copyberry
- #37204 Add durable user-message queue dispatch @copyberry
- #37206 Add a unified image budget @copyberry
- #37210 Fetch remote installed plugins across all scopes @copyberry
- #37211 Harden network proxy MITM authorization @copyberry
- #37248 Cover remote MCP discovery timeout cleanup @copyberry
- #37252 Allow agent roles on full-history forks @copyberry
- #37260 Fix first-turn model switching and rollback @copyberry
- #37261 Start cached MCP servers lazily for subagents @copyberry
- #37267 Support plugin roots in the host skill loader @copyberry
- #37273 Reuse MCP handlers across sampling steps @copyberry
- #37279 Avoid cloning immutable metadata on tool search cache hits @copyberry
- #37335 Fully repaint inline viewports after history overlap @copyberry
- #37336 Use step environments for extension turn input @copyberry
- #37337 Recover MCP servers after OAuth reauthentication @copyberry
- #37338 Honor the configured ChatGPT origin in connector install URLs @copyberry
- #37339 Reload app-server telemetry after account changes @copyberry
- #37340 Consolidate deferred environment provisioning APIs @copyberry
- #37341 Support content references for inline visualizations @copyberry
- #37342 Preserve foreign cwd URIs for turn-input contributors @copyberry
- #37343 Stage Bazel app-server test binaries in
TEST_TMPDIR@copyberry - #37344 Fix subagent MCP startup status settling @copyberry
- #37345 Send model routing hints to the Codex backend @copyberry
- #37347 Track context windows per agent @copyberry
- #37348 Add rollout migration tooling and background migration @copyberry
- #37349 Mount a minimal
/devin full-filesystem Bubblewrap sandboxes @copyberry - #37350 Allow
ThreadManagerto customize thread ID generation @copyberry - #37352 Configure the default code-mode exec yield timeout @copyberry
- #37354 Retry busy app-server test executable spawns @copyberry
- #37356 Support agent identity endpoint overrides @copyberry
- #37357 Clamp short wait_agent timeouts to the configured minimum @copyberry
- #37358 Add Markdown conversation export to the TUI @copyberry
- #37360 Use consistent TUI input placeholders @copyberry
- #37363 Recognize MCP tool hook configurations @copyberry
- #37366 Harden local MCP server process tree cleanup @copyberry
- #37367 Add session forking to
codex exec@copyberry - #37368 Restore approval policy when resuming threads @copyberry
- #37369 Add session archiving to the resume picker @copyberry
- #37371 Restore archived sessions from the resume picker @copyberry
- #37389 Add tool namespace metadata configuration @copyberry
- #37400 Rename the tool registry metadata setting @copyberry
- #37406 Add a loader for executor-local config layers @copyberry
- #37408 Add executor-local config reads to the exec server @copyberry
- #37424 Cap project instructions across environments @copyberry
- #37433 Expose multi-agent versions in model/list @copyberry
- #37434 Add process diagnostics snapshots @copyberry
- #37439 Add shared skill root loading interfaces @copyberry
- #37440 Load plugin skill roots through the host skills service @copyberry
- #37444 Unify plugin skill loading with the host skill service @copyberry
- #37446 Preserve base instruction provenance across sessions @copyberry
- #37447 Respect plugin skill availability in tool suggestions @copyberry
- #37452 Unify plugin skill loading through the shared loader @copyberry
- #37457 Remove the legacy core skill loader @copyberry
- #37461 Remove the unused remote skills client @copyberry
- #37466 Move skill config rule resolution into
codex-config@copyberry - #37470 Expose app-server diagnostics through the experimental API @copyberry
- #37477 Include call IDs in MCP requests and clarify metadata config @copyberry
- #37478 Shard state unit tests under Bazel @copyberry
- #37479 Report temporary directories in exec-server environment info @copyberry
- #37480 Delegate remote process sandboxing to the executor @copyberry
- #37483 Interrupt active code-mode cells with their turn @copyberry
- #37485 Keep response streams alive through connection failures @copyberry
- #37486 Expose runtime activity in server diagnostics @copyberry
- #37488 Generalize skill locator aliases across providers @copyberry
- #37489 Alias resource-backed skill locators under context pressure @copyberry
- #37492 Include tool namespace inventory in turn metadata @copyberry
- #37494 Add MCP event discovery and subscriptions @copyberry
- #37497 Limit payload traces in diagnostic logs @copyberry
- #37498 Preserve child waiters during process termination @copyberry
- #37500 Remove the legacy code-mode tool metadata inventory @copyberry
- #37503 Move host skill prompt injection into the skills extension @copyberry
- #37504 Disable Nagle's algorithm for code-mode WebSockets @copyberry
- #37505 Remove the codex-core-skills crate @copyberry
- #37507 Include sandbox mode in response metadata @copyberry
- #37510 Define the code-mode host gRPC protocol @copyberry
- #37511 Enforce automatic review for managed models @copyberry
- #37513 Reuse parent compactions in Guardian review sessions @copyberry
- #37516 Ignore reusable command approvals for cyber models @copyberry
- #37519 Expose auto-review ignore rules in config requirements @copyberry
- #37527 Terminate timed-out hook process trees @copyberry
- #37528 Keep external agent detection from blocking config requests @copyberry
- #37530 Implement the gRPC code-mode host service @copyberry
- #37533 Support asynchronous command hooks @copyberry
- #37538 Expose execution mode in hook listings @copyberry
- #37607 Prevent launch context from reaching child processes @copyberry
- #37610 Add workload identity token exchange support @copyberry
- #37618 Use step environments for Guardian approval reviews @copyberry
- #37622 Include buffered turns when editing prompts @copyberry
- #37641 Use the step context for command approval prefix rules @copyberry
- #37644 Generalize hook handler execution @copyberry
- #37645 Improve plugin install failure analytics @copyberry
- #37654 Advertise environment config read support @copyberry
- #37709 Keep wrapped composer whitespace with following text @copyberry
- #37723 Report I/O subtypes for session config import failures @copyberry
- #37745 Add gRPC TCP transport to the code-mode host @copyberry
- #37747 Bound Cursor project path resolution @copyberry
- #37757 Add a line-ending preservation mode to
apply_patch@copyberry - #37758 Add a feature flag to preserve apply_patch line endings @copyberry
- #37773 Forward install attempt IDs for remote plugins @copyberry
- #37788 Use native transparency in the imagegen skill @copyberry
- #37806 Initialize the install attempt ID in the plugin analytics test @copyberry
- #37807 Share model-visible tool specs across prompts @copyberry
- #37808 Simplify package-based skill reads @copyberry
- #37810 Support packaged defaults in config layering @copyberry
- #37812 Keep multi-workspace skill listings consistent @copyberry
- #37828 Track running unified exec processes at turn completion @copyberry
- #37832 Remove obsolete plugin skill discovery helpers @copyberry
- #37833 Encapsulate watchable skill root selection @copyberry
- #37836 Run plugin and skill tests on Windows @copyberry
- #37838 Tighten the skills extension API surface @copyberry
- #37842 Speed up MCP OAuth credential reads @copyberry
- #37843 Use the shared environment scrubber in git-utils @copyberry
- #37847 Preserve environments when reloading V2 agents @copyberry
- #37848 Expose the session ID to shell commands @copyberry
- #37850 Expose plugin ownership in MCP server status @copyberry
- #37851 Route intercepted exec approvals through shared review @copyberry
- #37860 Speed up MCP OAuth credential reads @copyberry
- #37862 Rename environment config for turn scope @copyberry
- #37864 Support MCP form input in full-access user threads @copyberry
- #37866 Add MCP OAuth credential contention regression tests @copyberry
- #37867 Reject duplicate resolved paths in apply_patch @copyberry
- #37871 Extract persisted history types into a dedicated crate @copyberry
- #37874 Keep runtime summary metrics out of Statsig exports @copyberry
- #37875 Honor the configured Windows sandbox level for managed networking @copyberry
- #37878 Add configurable goal token budget limits @copyberry
- #37882 Read safety buffering from response metadata @copyberry
- #37886 Extend bundled package discovery and expose its version @copyberry
- #37889 Ignore Unix socket proxy settings on Windows @copyberry
- #37891 Use thread configuration for
app/read@copyberry - #37892 Validate images before returning
view_imageoutput @copyberry - #37895 Add configurable Responses API request metadata @copyberry
- #37896 Add hermetic Windows SDK and MSVC runtime repositories @copyberry
- #37898 Add appearance metadata to thread sections @copyberry
- #37901 Make submission operations move-only @copyberry
- #37902 Defer
view_imageprocessing to history insertion @copyberry - #37906 Make gRPC code-mode notifications fire-and-forget @copyberry
- #37908 Apply refreshed cloud config bundles to later sessions @copyberry
- #37922 Extract reusable code-mode host test support @copyberry
- #37926 Distinguish turn-start thread persistence @copyberry
- #37929 Add shared runtime build information @copyberry
- #37939 Validate images before returning
view_imageoutput @copyberry - #37970 Cache tool catalogs for streamable HTTP MCP servers @copyberry
- #37979 Honor per-directory bundled skill settings in
skills/list@copyberry - #37984 Stop re-exporting skill APIs from
codex-core@copyberry - #38011 Remove config lockfile support @copyberry
- #38020 Retry transient exec-server startup failures @copyberry
- #38024 Expose image generation usage-limit failures @copyberry
- #38026 Fail closed on unsafe Linux unreadable globs @copyberry
- #38032 Hide approved Guardian assessments from TUI history @copyberry
- #38033 Use session metadata to validate thread history paths @copyberry
- #38034 Filter live rollout items in place @copyberry
- #38035 Propagate MCP elicitation event delivery failures @copyberry
- #38036 Limit TUI streaming traces in SQLite logs @copyberry
- #38040 Propagate custom CA settings to local MCP servers @copyberry
- #38041 Add gRPC-backed code-mode sessions @copyberry
- #38043 Sandbox remote apply_patch operations @copyberry
- #38044 Compact code mode tool calls in TUI history @copyberry
- #38045 Store model history in response item envelopes @copyberry
- #38046 Include auto-review state in turn metadata @copyberry
- #38047 Add turn-aware response item injection @copyberry
- #38049 Harden network proxy credential brokerage @copyberry
- #38051 Run required CI against pull request merge commits @copyberry
- #38052 Add per-login MCP OAuth client registration selection @copyberry
- #38054 Add configuration-backed external authentication @copyberry
- #38056 Configure PSP routing through the feature system @copyberry
- #38057 Track artifact operations from trusted plugin markers @copyberry
- #38058 Preserve harness metadata across conversation history @copyberry
- #38060 Disable storage for Azure Responses requests @copyberry
- #38061 Preserve proxy settings for Windows sandbox debug sessions @copyberry
- #38064 Grant Windows sandbox access to the Codex app root @copyberry
- #38066 Track resource-backed skill invocations @copyberry
- #38067 Scope environment readiness config to thread attachments @copyberry
- #38072 Forward gRPC code-mode callbacks to session delegates @copyberry
- #38074 Track implicit executor skill invocations @copyberry
- #38075 Respect rendered width when adding TUI history @copyberry
- #38078 Reduce cloning in world-state patch handling @copyberry
- #38080 Allow nested Git repositories in the Windows sandbox @copyberry
- #38081 Use
ReviewDecisionfor MCP tool approvals @copyberry - #38083 Remove standard form input from app-server docs @copyberry
- #38084 Allow empty input to start a turn @copyberry
- #38086 Support execution-host context when resolving cloud config @copyberry
- #38087 Route gRPC code-mode sessions through the shared HTTP client @copyberry
- #38089 Add CIMD support to MCP OAuth registration @copyberry
- #38092 Simplify queued user message admission @copyberry
- #38094 Test Guardian context for code mode commands @copyberry
- #38101 Attach hosted app context to file uploads @copyberry
- #38103 Avoid cloning MCP invocations in TUI history @copyberry
- #38108 Route MCP tool calls through shared approval handling @copyberry
- #38127 Distinguish rollout IDs from thread IDs @copyberry
- #38165 Resolve subagent analytics connections lazily @copyberry
- #38167 Read executor skill packages directly @copyberry
- #38170 Notify running turn watchers only on count changes @copyberry
- #38172 Update
lruandwebbrowserdependencies @copyberry - #38179 Embed defaults in the packaged config layer @copyberry
- #38183 Add conservative restriction helpers for tool policies @copyberry
- #38184 Run search tool integration tests on Windows @copyberry
- #38186 Stop overriding environments in the skills user-turn test @copyberry
- #38188 Integrate workload identity with Codex authentication @copyberry
- #38189 Run plugin app-server tests in automatic environments @copyberry
- #38197 Add an LRU baseline to skill shadow selection @copyberry
- #38204 Fuse recent and lexical skills in shadow selection @copyberry
- #38205 Enforce non-interactive approval policy for Codex delegates @copyberry
- #38214 Avoid allocations when sanitizing TUI user text @copyberry
- #38217 Start required cached MCP servers lazily for subagents @copyberry
- #38227 Add a flag to retain client developer messages @copyberry
- #38228 Detect implicit skill invocations from PowerShell reads @copyberry
- #38232 Track root turns across delegated Codex requests @copyberry
- #38238 Add manifest-defined metrics for trusted plugin scripts @copyberry
- #38239 Add bounded plugin measurement analytics @copyberry
- #38241 Include Node REPL policy in turn metadata @copyberry
- #38242 Cache stable active-cell layout measurements @copyberry
- #38243 Track client-authored developer messages in rollout history @copyberry
- #38244 Resolve paginated thread history by rollout ID @copyberry
- #38245 Add dynamic HTTP header helpers for MCP servers @copyberry
- #38251 Read model ETags from WebSocket metadata events @copyberry
- #38252 Collect metrics from plugin shell commands @copyberry
- #38253 Collect plugin metrics from unified exec commands @copyberry
- #38256 Report the latest rejection from multiple network reviews @copyberry
- #38257 Reconnect gRPC code-mode sessions after host restarts @copyberry
- #38258 Unify external authentication provider handling @copyberry
- #38261 Resolve skill package aliases in
skills.read@copyberry - #38265 Use bounded fallback ports for Windows managed proxies @copyberry
- #38268 Expose executor skill roots from
skills.read@copyberry - #38270 Add per-thread usage queries to the backend client @copyberry
- #38272 Stamp conversation history items with creation times @copyberry
- #38274 Represent persisted world state as JSON objects @copyberry
- #38275 Unify turn input submission and routing @copyberry
- #38276 Track plugin metrics for background unified exec commands @copyberry
- #38278 Add app-server coverage for plugin measurement analytics @copyberry
- #38280 Use protobuf's built-in Bazel proto rule @copyberry
- #38281 Show estimated thread usage in
/status@copyberry - #38282 Add thread usage to TUI status surfaces @copyberry
- #38283 Collect plugin metrics from remote executors @copyberry
- #38285 Move
codex-execpolicyto protocol dev dependencies @copyberry - #38288 Support gRPC code-mode hosts in app server @copyberry
- #38291 Remove unused apply_patch prompt fallback @copyberry
- #38292 Add durable reverts for paginated threads @copyberry
- #38299 Route network access through the shared approval pipeline @copyberry
- #38303 Add interrupted turn recovery @copyberry
- #38306 Protect inline visualization viewers from sandbox writes @copyberry
- #38321 Make gRPC code-mode yield tests deterministic @copyberry
- #38336 Add Guardian V2 extension scaffold @copyberry
- #38356 Support sandboxed file streaming in exec-server @copyberry
- #38358 Optimize orphan output normalization @copyberry
- #38361 Test hook rejection for explicitly started queue items @copyberry
- #38362 Stabilize exec-server byte-budget tests @copyberry
- #38363 Persist security risk scores in rollout history @copyberry
- #38368 Add the Guardian V2 Luna sampler @copyberry
- #38377 Constrain Guardian reviews to parent filesystem permissions @copyberry
- #38380 Preserve user message styling when wrapping long URLs @copyberry
- #38381 Prevent unread events from blocking in-process requests @copyberry
- #38383 Return Luna samples when streamed JSON completes @copyberry
- #38384 Refine skill creation guidance and validation @copyberry
- #38390 Use effective permissions when trusting app-server projects @copyberry
- #38394 Reject sessions with unloadable required managed hooks @copyberry
- #38396 Reap orphaned processes in Linux sandboxes @copyberry
- #38397 Include Node REPL results in Guardian reviews @copyberry
- #38399 Preserve floating-point values when decoding rollout lines @copyberry
- #38400 Refresh tracing interest in the token estimate test @copyberry
- #38403 Expose conversation history to tool lifecycle extensions @copyberry
- #38406 Pool Guardian sampling WebSocket connections @copyberry
- #38409 Classify tool calls in the Guardian V2 extension @copyberry
- #38413 Allow metadata updates without materializing threads @copyberry
- #38414 Add bounded transcript rendering for Guardian v2 @copyberry
- #38415 Recognize PowerShell
Get-Contentfile reads @copyberry - #38416 Honor filesystem permissions for app file uploads @copyberry
- #38418 Clarify MCP OAuth reauthentication errors @copyberry
- #38419 Apply Codex attribution to app-created commits @copyberry
- #38420 Recover capability discovery after executor disconnects @copyberry
- #38423 Centralize thread environment selection state @copyberry
- #38424 Fail closed when workload identity initialization fails @copyberry
- #38426 Protect workload identity auth in app-server account RPCs @copyberry
- #38427 Add Guardian guidance for Node REPL tool calls @copyberry
- #38429 Route curated plugin catalogs by authentication mode @copyberry
- #38432 Gate Node REPL Guardian guidance on model metadata @copyberry
- #38436 Add rustls fallback for local MCP HTTP requests @copyberry
- #38439 Preserve approval policies for auto-reviewed models @copyberry
- #38440 Add app-server support for reverting paginated threads @copyberry
- #38441 Give Guardian V2 full tool action context @copyberry
- #38443 Tag current time reminders in model context @copyberry
- #38445 Retain client developer messages across context compaction @copyberry
- #38446 Refresh current-time reminders for full-history subagents @copyberry
- #38447 Add running-task exit choices to local daemon sessions @copyberry
- #38448 Support per-server MCP OAuth callback ports @copyberry
- #38449 Expose model upgrade retirement times @copyberry
- #38450 Embed the Windows sandbox setup manifest in Bazel builds @copyberry
- #38452 Add structured telemetry for response retries @copyberry
- #38454 Include node_repl images in Guardian review evidence @copyberry
- #38456 Add experimental thread queue APIs to app server @copyberry
- #38460 Add an
AbsolutePathBufconversion forFileSystemPath@copyberry - #38461 Centralize turn environment selection state @copyberry
- #38463 Preserve thread subscriptions across revert reloads @copyberry
- #38467 Parse model annotations from skill frontmatter @copyberry
- #38470 Add an Amazon Bedrock Runtime provider @copyberry
- #38473 Stop generating accepted-line fingerprints @copyberry
- #38475 Add bounded skill model delegation instructions @copyberry
- #38483 Include agent names in turn metadata @copyberry
- #38484 Search selected plugin apps before falling back @copyberry
- #38492 Enforce strict auto-review for MCP tool calls @copyberry
- #38493 Reuse pending MCP startups during reconciliation @copyberry
- #38494 Verify bundled bwrap in Bazel builds @copyberry
- #38499 Enable parallel tool calls for all model prompts @copyberry
- #38521 Carry environment config in turn selections @copyberry
- #38527 Configure Guardian sampling for responses lite @copyberry
- #38540 Record Guardian V2 risk scores on threads @copyberry
- #38553 Initialize Guardian V2 samplers per thread @copyberry
- #38555 Preserve Guardian transcript boundaries in sampling input @copyberry
- #38558 Improve Guardian transcript context selection @copyberry
- #38561 Restrict filesystem helper sandbox access @copyberry
- #38564 Reuse compatible parent compactions in Guardian V2 @copyberry
- #38567 Store security risk scores as a snapshot @copyberry
- #38568 Run tool start callbacks after pre-tool hooks @copyberry
- #38569 Require automatic review for high-risk Guardian v2 actions @copyberry
- #38579 Report thread storage changes after rollout migration @copyberry
- #38580 Keep the latest Guardian risk score during concurrent sampling @copyberry
- #38581 Honor cloud-managed requirements in feature listings @copyberry
- #38586 Bound Guardian V2 tool actions before risk classification @copyberry
- #38592 Let extensions resolve approval reviews before Guardian @copyberry
- #38596 Prioritize new Guardian classifications under load @copyberry
- #38597 Install Guardian V2 in the app server @copyberry
- #38601 Make unbounded connection retries configurable @copyberry
- #38602 Isolate Guardian reviewer sessions from parent extensions @copyberry
- #38604 Avoid paginated resume requests for verified legacy rollouts @copyberry
- #38605 Fix resume transcript preview line selection @copyberry
- #38606 Allow larger gRPC code mode tool descriptions @copyberry
- #38608 Fix the missing resume picker history mode import @copyberry
- #38610 Support workload identity in remote exec-server auth @copyberry
- #38612 Skip stale Guardian risk score persistence @copyberry
- #38615 Remove the gRPC code-mode enabled tool limit @copyberry
- #38616 Route escalated approval retries through Guardian @copyberry
- #38617 Prioritize global models in the Bedrock Runtime catalog @copyberry
- #38618 Apply Guardian policies to v2 risk classification @copyberry
- #38619 Source multi-agent instructions from the model catalog @copyberry
- #38621 Preserve large gRPC code-mode tool errors @copyberry
- #38623 Preserve MCP namespace descriptions in the tool catalog cache @copyberry
- #38624 Open notes when accepting the request input Other option @copyberry
- #38625 Enable unified exec by default on Windows @copyberry
- #38628 Make Guardian v2 risk classification configurable @copyberry
- #38630 Remove the gRPC code-mode open session limit @copyberry
- #38634 Add MCP protocol discovery metrics @copyberry
- #38635 Remove repository-local Codex skills @copyberry
- #38639 Render the initial TUI session header before input @copyberry
- #38641 Harden TUI startup input handling @copyberry
- #38642 Keep the composer editable during TUI startup @copyberry
- #38643 Delay the startup composer until first-login onboarding @copyberry
- #38644 Show onboarding when Codex home lacks authentication state @copyberry
- #38645 Deliver gRPC code-mode notifications without truncation @copyberry
- #38646 Read Apple notarization issuer ID from Key Vault @copyberry
- #38647 Add an override to skip project configuration @copyberry
- #38649 Reuse the TUI startup account response during bootstrap @copyberry
- #38650 Canonicalize default namespaces in gRPC subscription filters @copyberry
- #38651 Move permission profile snapshots into the protocol @copyberry
- #38657 Skip terminal hyperlink layout when no links are present @copyberry
- #38660 Enforce managed deny-read rules in the Windows sandbox @copyberry
- #38662 Delete Thai combining marks one at a time in the composer @copyberry
- #38664 Resolve local JSON Schema refs in Code Mode types @copyberry
- #38670 Forward executor network policy decisions for auditing @copyberry
- #38673 Honor per-environment permission profiles @copyberry
- #38675 Exclude shortcut-modified input from TUI paste bursts @copyberry
- #38678 Preserve environment configuration ownership @copyberry
- #38681 Preserve HTTP fallback for delegated sessions @copyberry
- #38682 Surface misalignment policy violations as typed errors @copyberry
- #38684 Support pending environment attachment configuration @copyberry
- #38690 Propagate request trace context through exec-server relays @copyberry
- #38701 Route permission requests through shared Guardian approvals @copyberry
- #38703 Refresh hook runtimes after plugin changes @copyberry
- #38704 Normalize CRLF line endings in pasted text @copyberry
- #38705 Add MCP tool handler support to the hooks engine @copyberry
- #38743 Scope TUI app directory state to the active context @copyberry
- #38746 Remove the TUI test dependency on
codex-cli@copyberry - #38767 Forward workload identity context during token exchange @copyberry
- #38774 Use paginated history for persistent exec threads @copyberry
- #38785 Keep active-turn model settings stable across updates @copyberry
- #38788 Show resume and fork status during TUI startup @copyberry
- #38795 Add storage diagnostics to
codex doctor@copyberry - #38800 Route executor policy audits through log-only telemetry @copyberry
- #38806 Add a health endpoint to the code-mode gRPC listener @copyberry
- Export complete TUI conversations to Markdown with
-
Public plugin catalog CSV export
Eligible ChatGPT Enterprise workspace owners and admins can download a CSV of the public plugins visible to their workspace. The export includes plugin, app, and Chat skill names and descriptions, along with developer, version, date added in UTC, and OpenAI verification metadata.
Open Admin > Plugins, select Public, and then select the download icon (Export CSV). The export uses a public-catalog snapshot that can be up to 48 hours old and does not include plugins created for the workspace. It isn’t available in FedRAMP workspaces.
Learn more about plugin controls.
-
Computer History
Computer History is an opt-in feature in the ChatGPT desktop app on macOS that turns activity across apps and websites into memories and a timeline that ChatGPT and Codex can use. Choose which apps and websites contribute, pause collection, and review or delete your history at any time.
Computer History is available to ChatGPT Pro, Business, and Enterprise users. Business and Enterprise administrators must enable access before workspace members can turn it on. Initial availability excludes the European Economic Area (EEA), Switzerland, and the United Kingdom.
-
Linux desktop preview and agent imports
Install the ChatGPT desktop app on Linux
The ChatGPT desktop app for Linux is available in preview for supported Ubuntu, Debian, and Fedora desktop distributions on x64 and ARM64 processors. Download the
.debor.rpmpackage for your distribution, then sign in to work with projects, local files, and Codex.Import setup and recent work from other agents
The desktop app supports Claude Code, Claude Cowork, and Cursor. Import instructions, settings, skills, plugins, projects, and recent work, then turn on automatic updates in Settings > Import to keep imported work in sync.
Codex CLI can also import supported setup and recent chats from Claude Code and Cursor with
/import. -
Introducing Daybreak Blue and Daybreak Red
Daybreak now offers two access tiers for approved defenders: Daybreak Blue and Daybreak Red. Use them to move from security findings to validated fixes in explicitly authorized engagements.
Start with Daybreak Blue for most defensive security work. It provides access to general-purpose models such as GPT-5.6 Sol for vulnerability discovery, secure code review, detection engineering, incident response, malware analysis, and patch validation.
Daybreak Red provides separately approved access to purpose-trained models such as GPT-5.6 Cyber for authorized vulnerability reproduction, exploit validation, penetration testing, red teaming, and complex system analysis.
Access requires approval through Trusted Access for Cyber and applies only to the approved identity, workspace or API organization and project, model, and product surface. Daybreak Red requires separate approval and provisioning; Daybreak Blue access doesn’t grant access to Daybreak Red.
Work in an isolated environment, define your engagement scope, use least-privilege permission profiles, and configure Auto-review for eligible actions before they cross the sandbox boundary.
Learn more about Cyber Safety, Codex credit rates, and API token prices.
-
ChatGPT for iOS 1.2026.209
Improvements and bug fixes
- Improved task reconnection with visible progress while keeping existing tasks usable during connection checks.
- New tasks now preserve their first prompt and attachments while starting, preventing empty tasks that could not be reopened.
- Improved editing performance for long prompts.
- Voice conversations now follow your Background Conversations setting.
- Approvals and user-input requests now recover reliably after reconnecting, with cleaner approval details in the transcript.
- Fixed follow-up
/goalprompts after completing a goal. - Fixed task-list freezes, disappearing task titles, flashing thinking summaries, and transcript crashes.
-
Codex CLI 0.147.0
$ npm install -g @openai/codex@0.147.0View details
New Features
- Install portable Agent Plugins and search across local, personal, workspace, and remote plugin catalogs. (#36544, #36409, #36919, #36796)
- Organize conversations into persistent, manually ordered sections and browse long transcripts incrementally. (#35722, #36007, #36380, #36948, #36950)
- Enable automatically reviewed approvals with the new
--approve-for-meCLI flag. (#36373) - Import Cursor-managed skills and synchronize changes to imported Claude and Cursor conversations without creating duplicates. (#36361, #36356, #35623)
- Support the opt-in MCP 2026-07-28 protocol, including paginated discovery, multi-round requests, and non-blocking server startup. (#35724, #35725, #35590, #35742)
- Enable cached web search and remote conversation compaction for Amazon Bedrock. (#36938, #36981)
Bug Fixes
- Redact secrets and complete bearer tokens from displayed commands and replayed conversation history. (#36893, #36908)
- Prevent lost or stalled terminal input when focus returns, MCP servers initialize, or Ghostty handles keyboard shortcuts. (#35649, #35957, #36834)
- Correct rendering and cursor positioning for Japanese characters, emoji, hyperlinks, and text near viewport boundaries. (#35960, #35962, #37166)
- Properly interrupt Windows background processes and handle Windows filesystem paths consistently. (#35655, #35851, #37129)
- Require explicit trust for unfamiliar local projects and enforce managed authentication restrictions before credentials are used. (#36960, #37132)
- Harden plugin isolation and deny network access when policy updates fail. (#37027, #36967, #36037)
Documentation
- Improve the bundled OpenAI documentation skill with targeted official-source lookup and clearer guidance for Codex, model selection, and API workflows. (#36014)
Chores
- Upgrade the MCP SDK to 3.0.0, Ratatui to 0.30.2, and V8 to 150.4.0. (#36001, #35959, #35831)
- Secure macOS release notarization using Azure Key Vault instead of exporting private signing keys. (#37154)
- Remove the deprecated
codex exec --full-autoflag; use--sandbox workspace-writeinstead. (#36054) - Stop publishing redundant Linux bundle archives; use the standard
codex-package-<target>release archives. (#36342)
Changelog
Full Changelog: rust-v0.146.0...rust-v0.147.0
- #35590 Expose cached MCP tools before server startup @copyberry
- #35594 Recommend longer waits in the v2 wait_agent schema @copyberry
- #35595 Allow alpha hotfix versions in R2 releases @copyberry
- #35597 Add metrics for extension-rendered skill catalogs @copyberry
- #35607 Record metrics for empty skill catalogs @copyberry
- #35608 Support model-owned token budget defaults @copyberry
- #35621 Skip restored token usage replay for exec resumes @copyberry
- #35623 Parse Claude and Cursor session records separately @copyberry
- #35642 Make OpenTelemetry provider shutdown idempotent @copyberry
- #35644 Preserve thread metadata when rollout files are missing @copyberry
- #35649 Preserve TUI input when terminal focus returns @copyberry
- #35652 Enable network policy callbacks for remote exec @copyberry
- #35653 Test developer instruction inheritance for multi-agent workers @copyberry
- #35655 Terminate Windows non-TTY processes on interrupt @copyberry
- #35656 Preserve multi-agent settings across config representations @copyberry
- #35661 Place host skills before permission instructions @copyberry
- #35663 Evaluate character matching over skill routing metadata @copyberry
- #35665 Fix the async watcher test harness on Windows @copyberry
- #35668 Expose the network proxy spec constructor @copyberry
- #35670 Raise the Windows exec yield floor to 10 seconds @copyberry
- #35671 Route curated plugins by authentication mode @copyberry
- #35675 Prepare MCP and plugin recommendations concurrently @copyberry
- #35678 Preserve paginated thread metadata across resumes @copyberry
- #35685 Load cloud-managed profiles for
codex sandbox@copyberry - #35688 Point crossterm patch to the OpenAI OSS fork @copyberry
- #35689 Preserve item timestamps in thread history projections @copyberry
- #35691 Include empty-preview threads in relationship listings @copyberry
- #35693 Refresh the subagent picker in the background @copyberry
- #35695 Honor the configured SQLite home in the logs client @copyberry
- #35708 Add configurable developer instructions for v2 subagents @copyberry
- #35715 Support streaming bodies in route-aware HTTP requests @copyberry
- #35717 Reuse route-aware clients for OpenAI file uploads @copyberry
- #35720 Upgrade rmcp to 3.0.0-beta.3 @copyberry
- #35721 Honor disabled code mode fallback without a process host @copyberry
- #35722 Add persisted sections for organizing threads @copyberry
- #35724 Add MCP 2026-07-28 discovery support @copyberry
- #35725 Complete MCP 2026 client support @copyberry
- #35738 Add bounded metadata for executed tool calls @copyberry
- #35742 Avoid blocking turns on optional MCP startup @copyberry
- #35744 Keep agent registry identities consistent @copyberry
- #35766 Emit host skill budget warnings from world state @copyberry
- #35769 Share the skills budget across host and executor catalogs @copyberry
- #35772 Throttle models cache TTL renewals @copyberry
- #35773 Scale skill metadata budgets with context windows @copyberry
- #35777 Resolve MCP tool catalogs concurrently @copyberry
- #35779 Load thread titles concurrently during session startup @copyberry
- #35785 Support self-serve Business ProLite accounts @copyberry
- #35787 Gate paginated thread history on the state database @copyberry
- #35794 Wait for MCP readiness in the curated sync test @copyberry
- #35802 Tag reports with the selected turn's model and effort @copyberry
- #35806 Route MCP OAuth through configured HTTP clients @copyberry
- #35814 Use configured HTTP clients for all MCP OAuth requests @copyberry
- #35818 Advance latest-alpha-cli after release publishing completes @copyberry
- #35821 Use the shared HTTP client for TUI network checks @copyberry
- #35825 Use the shared HTTP client for announcement tips @copyberry
- #35828 Enforce centralized SQLite connection creation @copyberry
- #35830 Route WebRTC sideband joins to the Realtime API @copyberry
- #35831 Update rusty_v8 to 150.4.0 @copyberry
- #35835 Track parent turns for nested Codex requests @copyberry
- #35836 Clean up cancelled MCP elicitation requests @copyberry
- #35837 Expose plugin eligibility metadata in app-server summaries @copyberry
- #35839 Decouple recommended plugins from tool suggestions @copyberry
- #35840 Handle legacy MCP discovery prevalidation errors @copyberry
- #35843 Tie remote exec servers to their parent stdin @copyberry
- #35845 Support plaintext collaboration tool messages @copyberry
- #35850 Preserve foreign paths in background terminal listings @copyberry
- #35851 Normalize Windows namespace paths in path URIs @copyberry
- #35854 Box app-server event payloads @copyberry
- #35856 Resolve imported connectors by MCP server name @copyberry
- #35857 Add Bazel unit test targets for Rust binaries @copyberry
- #35859 Expose plugin installation timestamps in app-server summaries @copyberry
- #35870 Include session titles in external agent import history @copyberry
- #35874 Mark the primary environment in model context @copyberry
- #35875 Allow environment readiness updates in place @copyberry
- #35878 Use step environments for MCP file uploads @copyberry
- #35886 Show blocked goals as stalled in the TUI @copyberry
- #35887 Clean up side conversations in the background @copyberry
- #35895 Inherit ready step environments when spawning agents @copyberry
- #35937 Let unrelated tools run while MCP servers start @copyberry
- #35941 Bound MCP namespace descriptions @copyberry
- #35942 Update tests for current tool call and app-server event types @copyberry
- #35944 Report direct input capability for listed subagents @copyberry
- #35957 Fix TUI input queue handling during MCP startup @copyberry
- #35959 Upgrade Ratatui to 0.30.2 @copyberry
- #35960 Preserve hyperlink cell widths during terminal diffing @copyberry
- #35962 Fix TUI layout for halfwidth Japanese sound marks @copyberry
- #35982 Avoid cloning rollout history when truncating forks @copyberry
- #35989 Display title-only reasoning summaries in the TUI @copyberry
- #35990 Test exec-server compatibility across Codex versions @copyberry
- #35992 Group external agent config migration into a module @copyberry
- #35997 Remove obsolete rusty_v8 146.4.0 Bazel targets @copyberry
- #36001 Upgrade rmcp to 3.0.0 @copyberry
- #36002 Resolve MCP file uploads with environment-native paths @copyberry
- #36006 Reduce response serialization and rollout scan overhead @copyberry
- #36007 Add persisted manual ordering for thread sections @copyberry
- #36008 Route pet asset downloads through the shared HTTP client @copyberry
- #36011 Share optional MCP startup grace across connection sets @copyberry
- #36014 Refine OpenAI docs skill source routing @copyberry
- #36020 Avoid cloning ignored analytics notifications @copyberry
- #36030 Increase the app-server tracing test stack size @copyberry
- #36031 Load cloud-managed servers in MCP CLI commands @copyberry
- #36033 Use the shared HTTP client in codex-protocol @copyberry
- #36035 Exit the stdio app-server when its connection closes @copyberry
- #36036 Allow naming forked chats from the TUI @copyberry
- #36037 Deny network access when an allow amendment fails @copyberry
- #36039 Limit MCP catalog pagination @copyberry
- #36043 Document the Responses API proxy reqwest exception @copyberry
- #36045 Distinguish unknown MCP authentication status @copyberry
- #36047 Extract MCP environment headers into a local variable @copyberry
- #36049 Keep tool-call metrics out of Statsig exports @copyberry
- #36051 Avoid overwriting symlinked migration targets @copyberry
- #36054 Remove legacy
--full-autohandling fromcodex exec@copyberry - #36055 Expose MCP read-only hints in tool call items @copyberry
- #36076 Remove codex-core's direct reqwest dependency @copyberry
- #36078 Route Ollama through the shared HTTP client @copyberry
- #36083 Handle fractional rate limits in thread history projection @copyberry
- #36092 Correlate image requests and explicit interrupts with turns @copyberry
- #36119 Unify tool runtime assembly before router construction @copyberry
- #36120 Delegate readiness waits to tool runtimes @copyberry
- #36121 Sandbox executor skill resource reads @copyberry
- #36124 Respect filesystem permissions during capability discovery @copyberry
- #36127 Centralize tool registration and protect host tools @copyberry
- #36128 Preserve delegated tasks across remote compaction @copyberry
- #36129 Resolve normalized tool name collisions in code mode @copyberry
- #36133 Preserve executor workspace permissions for apply_patch @copyberry
- #36138 Isolate apply-patch permission test variants @copyberry
- #36139 Bound the code mode output termination test @copyberry
- #36140 Wait for memory consolidation shutdown before finishing jobs @copyberry
- #36146 Stabilize approval scenario tests @copyberry
- #36148 Harden the file change approval integration test @copyberry
- #36149 Strengthen session file change approval coverage @copyberry
- #36165 Stabilize file approval replay test @copyberry
- #36166 Make pending realtime sideband test deterministic @copyberry
- #36167 Harden the auto-review model override test @copyberry
- #36168 Make tool telemetry tag collection synchronous @copyberry
- #36171 Test workspace-root patch protection for metadata @copyberry
- #36177 Test remote Guardian network approval scoping @copyberry
- #36181 Record attempted tool calls in response metadata @copyberry
- #36183 Use permission profiles throughout sandbox execution @copyberry
- #36184 Coalesce concurrent remote metadata requests @copyberry
- #36187 Refresh environment dates from the configured clock @copyberry
- #36188 Make thread history projection resilient to malformed rollouts @copyberry
- #36191 Label explicit skill injection metrics @copyberry
- #36194 Avoid shifting bytes in streaming output buffers @copyberry
- #36207 Record normalized sandbox violation events @copyberry
- #36212 Precompute app-server protocol exports @copyberry
- #36217 Run code mode exclusively through the standalone host @copyberry
- #36218 Expose connector candidates in external agent detection @copyberry
- #36221 Ignore passthrough metadata when reconciling rollout items @copyberry
- #36223 Preserve executor paths in read command actions @copyberry
- #36228 Support Enterprise automation account plans @copyberry
- #36237 Ignore symbolic slash-tmp permissions on Windows @copyberry
- #36239 Refresh precomputed app-server protocol exports @copyberry
- #36264 Prefer the bundled resource for the code mode host @copyberry
- #36271 Derive report prompt hashes from persisted rollouts @copyberry
- #36306 Restrict hosted MCP credentials to local environments @copyberry
- #36309 Use the host skill provider catalog in world state @copyberry
- #36310 Isolate MCP OAuth credentials by environment @copyberry
- #36311 Include skills in debug prompt input output @copyberry
- #36327 Preserve discovery order for host skill path aliases @copyberry
- #36329 Restrict
shell_commandto a single local environment @copyberry - #36336 Detect connectors used in external agent sessions @copyberry
- #36339 Enable skills in the MCP server @copyberry
- #36342 Stop publishing legacy Linux bundle archives @copyberry
- #36350 Require explicit sandbox permissions with shell justifications @copyberry
- #36351 Use model catalog collaboration mode messages @copyberry
- #36354 Pass sessions directly to session tasks @copyberry
- #36355 Keep MCP tool calls bound to their thread @copyberry
- #36356 Sync updates to imported external agent sessions @copyberry
- #36357 Use the step-scoped router for tool execution @copyberry
- #36359 Consolidate MCP config editing in codex-core @copyberry
- #36360 Use MCP bindings as the step tool catalog @copyberry
- #36361 Migrate Cursor-managed skills into Codex @copyberry
- #36364 Move skill catalog rendering out of core @copyberry
- #36365 Add strict automatic review for MCP elicitations @copyberry
- #36367 Keep effective tool exposure in the registry @copyberry
- #36372 Run native Windows Bazel tests with MSVC @copyberry
- #36373 Add an
--approve-for-meCLI flag @copyberry - #36374 Enable sandboxed V8 for code mode @copyberry
- #36378 Load local session pickers from the state DB first @copyberry
- #36380 Add thread section management APIs @copyberry
- #36384 Load turn summaries with paginated queries @copyberry
- #36385 Add acknowledged user message submission to core @copyberry
- #36388 Track image preparation details in turn analytics @copyberry
- #36389 Enforce single-writer ownership for all thread histories @copyberry
- #36393 Avoid redundant filesystem probes @copyberry
- #36402 Declare the experimental plugin search API @copyberry
- #36408 Allow custom Codex instructions for realtime transitions @copyberry
- #36409 Implement remote plugin search @copyberry
- #36410 Make user input blocking behavior explicit @copyberry
- #36411 Use Git repositories as pre-tool hook test markers @copyberry
- #36413 Add a realtime delegation acknowledgement control @copyberry
- #36440 Extract exec-server request dispatching @copyberry
- #36482 Avoid querying terminal size on every TUI redraw @copyberry
- #36485 Increase remote plugin bundle size limits @copyberry
- #36507 Retain attempted tool metadata across prompts @copyberry
- #36511 Support two-stroke TUI key chords @copyberry
- #36534 Raise the MCP catalog item limit to 2,048 @copyberry
- #36544 Support portable Agent Plugins throughout installation @copyberry
- #36632 Preserve SQLite thread metadata during goal mutations @copyberry
- #36635 Expose onboarding hints in login completion notifications @copyberry
- #36641 Capture rollout budget units from response usage @copyberry
- #36715 Use provider-reported rollout budget units @copyberry
- #36716 Deduplicate MCP resource operation handling @copyberry
- #36727 Simplify turn metadata state ownership @copyberry
- #36729 Correlate code mode tool analytics with model responses @copyberry
- #36732 Stabilize the cloud environment HTTP test @copyberry
- #36734 Deduplicate MCP resource list handling @copyberry
- #36740 Store turn skill state in extension data @copyberry
- #36742 Simplify contextual user fragment registration @copyberry
- #36744 Store the extension registry directly in its builder @copyberry
- #36745 Consolidate apply_patch runtime execution @copyberry
- #36757 Wait for patch events before reading shell snapshots @copyberry
- #36759 Harden the TUI focus palette regression test @copyberry
- #36764 Avoid building code-mode definitions during registration @copyberry
- #36769 Stabilize the startup Git enrichment integration test @copyberry
- #36770 Stabilize deferred goal fork testing @copyberry
- #36771 Harden Linux managed proxy helper lifecycles @copyberry
- #36772 Raise the host-owned Codex Apps catalog limit @copyberry
- #36773 Derive the active user config layer from the layer stack @copyberry
- #36774 Clarify config layer iteration APIs @copyberry
- #36779 Stabilize network policy event capture in concurrent tests @copyberry
- #36781 Add per-surface MCP tool exposure controls @copyberry
- #36782 Add Guardian session reuse and interruption tests @copyberry
- #36787 Consolidate model instructions in
ModelMessages@copyberry - #36792 Gate plugin usage instructions by model capability @copyberry
- #36793 Terminate timed-out Git process trees @copyberry
- #36796 Add Agent Plugins MCP config parsing @copyberry
- #36797 Normalize rusty_v8 checksum manifest line endings @copyberry
- #36800 Avoid reinjecting permissions after command approvals @copyberry
- #36807 Extract audio preparation into a utility crate @copyberry
- #36808 Prefer SQLite names for local session archive commands @copyberry
- #36809 Prefer the state database for
exec resume --last@copyberry - #36810 Add MCP client conformance regression gates @copyberry
- #36811 Honor per-environment login shell policy @copyberry
- #36812 Add a dual-WebSocket transport for code mode @copyberry
- #36815 Identify agents by name in token budget context @copyberry
- #36822 Fix typo in approval resolver name @copyberry
- #36825 Consolidate approval telemetry context @copyberry
- #36830 Time out stalled code-mode host requests @copyberry
- #36834 Avoid requesting key-release events in Ghostty @copyberry
- #36840 Keep API request metrics out of Statsig exports @copyberry
- #36856 Support deferred loading for freeform tools @copyberry
- #36857 Support custom tools in namespaces @copyberry
- #36862 Consolidate thread spawning behind a request object @copyberry
- #36871 Parallelize R2 asset publishing with DotSlash @copyberry
- #36877 Move executor skill bundle loading into the skills extension @copyberry
- #36880 Move direct executor skill discovery into the skills extension @copyberry
- #36882 Preserve complete MCP namespace descriptions @copyberry
- #36884 Add host skill root loading @copyberry
- #36892 Support leaf models in multi-agent v2 @copyberry
- #36893 Redact secrets from app-server command execution items @copyberry
- #36895 Handle late MCP startup results after lag timeout @copyberry
- #36898 Avoid redundant filesystem metadata probes @copyberry
- #36900 Register app tools independently of the connector list @copyberry
- #36901 Propagate updated permissions to review threads @copyberry
- #36903 Load skill interface metadata in the host loader @copyberry
- #36904 Limit RMCP logs persisted to SQLite @copyberry
- #36906 Preserve model providers when reloading v2 agents @copyberry
- #36908 Improve bearer token secret redaction @copyberry
- #36910 Negotiate MCP extensions per app-server session @copyberry
- #36912 Read approval policy from the current turn configuration @copyberry
- #36913 Move skill policy resolution into
codex-skills@copyberry - #36916 Centralize app enabled-state evaluation @copyberry
- #36917 Test explicit plugin mentions with disabled apps @copyberry
- #36919 Merge local plugins into plugin search results @copyberry
- #36921 Move the host skills service into the skills extension @copyberry
- #36924 Reject implicitly discovered bare Git repositories @copyberry
- #36930 Read turn permissions from the current configuration @copyberry
- #36935 Trust undecided local projects automatically @copyberry
- #36938 Enable cached web search for Amazon Bedrock @copyberry
- #36939 Include policy approval reasons in Guardian reviews @copyberry
- #36941 Use current session settings for review threads @copyberry
- #36943 Move host skill root resolution into the skills extension @copyberry
- #36945 Grant the blob size policy job read access to contents @copyberry
- #36947 Accept user input when starting idle turns @copyberry
- #36948 Paginate transcript history in the TUI @copyberry
- #36949 Improve paginated TUI history loading @copyberry
- #36950 Paginate TUI transcript history @copyberry
- #36951 Harden paginated history handling in the TUI @copyberry
- #36952 Add durable per-thread user submission queues @copyberry
- #36954 Add tool registry collision policy configuration @copyberry
- #36956 Keep image resize notices attached during remote compaction @copyberry
- #36959 Track connectors detected in external agent sessions @copyberry
- #36960 Prompt before trusting local project directories @copyberry
- #36963 Link Codex attribution in pull request bodies @copyberry
- #36964 Preserve working directories when importing external sessions @copyberry
- #36966 Allow disabling the built-in image viewer @copyberry
- #36967 Skip symlinks when installing plugins @copyberry
- #36970 Make token budget context identity configurable @copyberry
- #36976 Honor explicit-only orchestrator skills @copyberry
- #36977 Improve connector detection for migrated sessions @copyberry
- #36979 Add Fence auditing to the blob size workflow @copyberry
- #36981 Enable remote compaction for Amazon Bedrock @copyberry
- #36983 Preserve ChatGPT auth for trusted staging MCP servers @copyberry
- #36984 Support configured ChatGPT cookies in HTTP clients @copyberry
- #36986 Add process-scoped PSP routing for ChatGPT requests @copyberry
- #36987 Add opt-in concurrent exec-server request dispatch @copyberry
- #36989 Preserve shared bundled skill caches @copyberry
- #36990 Remove legacy collaboration mode variants @copyberry
- #36992 Allow injecting model catalog caches @copyberry
- #36993 Support
includeTurnsreads for paginated threads @copyberry - #36998 Support deferred custom tools in tool search @copyberry
- #37000 Keep shared skill caches fresh across plugin loads @copyberry
- #37020 Enforce strict tool name collision errors @copyberry
- #37022 Canonicalize default tools under the
functionsnamespace @copyberry - #37027 Enforce Agent Plugin runtime boundaries @copyberry
- #37031 Apply permission profile updates to future turn environments @copyberry
- #37035 Align registry tests with canonical tool names @copyberry
- #37038 Use turn environment permissions for tool execution @copyberry
- #37040 Use turn environment permissions for context and discovery @copyberry
- #37053 Reject conflicting namespace descriptions in strict tool mode @copyberry
- #37055 Add safer TUI defaults for cyber models @copyberry
- #37066 Preserve longer MCP source descriptions in tool search @copyberry
- #37068 Fall back to per-process MCP cleanup on macOS @copyberry
- #37083 Consolidate unified exec output state @copyberry
- #37101 Reuse stable MCP bindings across sampling steps @copyberry
- #37103 Use Luna for API-key Guardian reviews @copyberry
- #37109 Bound interactive telemetry shutdown @copyberry
- #37114 Add per-session code-mode execution limits @copyberry
- #37128 Centralize tool approval handling in
Session@copyberry - #37129 Make Windows path URI comparisons ASCII-case-insensitive @copyberry
- #37132 Enforce managed authentication requirements locally @copyberry
- #37134 Report prompt image resizing to the model @copyberry
- #37144 Preserve discovery paths for symlinked skills @copyberry
- #37145 Gate Apps usage instructions by model capability @copyberry
- #37147 Track provisioned environment state across registration @copyberry
- #37149 Project orchestrator skills through world state @copyberry
- #37151 Coalesce concurrent Git status scans @copyberry
- #37154 Use Azure Key Vault for macOS notarization @copyberry
- #37156 Test remote environments reported ready before selection @copyberry
- #37157 Harden named session lookup in the TUI @copyberry
- #37162 Load host skill roots through the skills extension @copyberry
- #37166 Keep textarea cursors and rendering inside the viewport @copyberry
- #37167 Expose session sources to MCP contributors @copyberry
- #37168 Bound remote MCP handshake HTTP requests @copyberry
- #37169 Move plugin skill snapshot integration tests into core @copyberry
- #37174 Centralize skill invocation helpers in
codex-skills@copyberry - #37175 Add legacy rollout migration to paginated history @copyberry
- #37177 Move explicit skill selection into the skills crate @copyberry
- #37178 Preserve image transparency metadata in app-server items @copyberry
- #37188 Reserve the
tool_searchnamespace for the search tool @copyberry - #37189 Track multi-agent usage hints in world state @copyberry
-
Codex CLI 0.146.1
$ npm install -g @openai/codex@0.146.1View details
Bug Fixes
- Apply safer automatic-review defaults for cyber-capable models and explain permission changes in the terminal interface. (#37057)
Changelog
Full Changelog: rust-v0.146.0...rust-v0.146.1
July 2026
-
GPT-5.4 and GPT-5.4 mini retire from Codex on August 31
On August 31, 2026, GPT-5.4 and GPT-5.4 mini will no longer be available in Codex for users signed in with ChatGPT. GPT-5.4 and GPT-5.4 mini will remain available on the OpenAI API and Codex sessions authenticated with an API key.
Switch to their recommended replacements:
- Replace
gpt-5.4withgpt-5.6-terra(GPT-5.6 Terra). - Replace
gpt-5.4-miniwithgpt-5.6-luna(GPT-5.6 Luna).
Before the cutoff, update workspace defaults, saved model settings, managed configurations, custom agents, and scheduled tasks that use either model.
See Codex models and workspace model availability for details.
- Replace
-
Record & Replay expands to the EU, UK, and Switzerland
Record & Replay is now available in the European Union, the United Kingdom, and Switzerland. On macOS, demonstrate a workflow and turn it into a reusable skill. Computer Use must also be available and enabled.
-
Browser upgrades, multi-repository review, and image editing 26.727
The latest ChatGPT desktop app update makes browsing, reviewing code, and editing generated images faster and easier.
Browse and find context faster
- Type in the built-in browser’s address bar to revisit pages from your browsing history or search Google when there’s no match.
- Manage your browsing history in Settings, and let ChatGPT search that history when a task needs to find a page you visited before.
- Use the Chrome extension to mention open tabs or bring highlighted page text into your side chat.
- Ask questions about any YouTube video in the Chrome extension and get answers in seconds.
- Right-click a webpage and select Ask ChatGPT.
Learn more about the built-in browser and the Chrome extension.
Review changes across repositories
See all repositories in a multi-folder project and the lines changed in each one. Select Review to inspect diffs across those repositories without switching between separate review views.
Refine generated images
Open generated images in an expanded viewer, and switch between Focused view and Canvas view. Add comments across images, choose the ones you want, and send targeted edits without leaving your conversation. Learn more about image generation.
Other improvements and bug fixes
- Added a new “Activity view” in the sidebar to view which chats you engaged with recently and require attention. Click the bell or use Cmd/Ctrl+Opt+U to change to the new view.
- Updated browser settings to show only supported browsers.
- Improved Windows installation reliability when package file paths are long.
- Other performance and bug fixes.
-
Sign in with ChatGPT (beta)
Sign in with ChatGPT is beginning to roll out in beta across select plugins and partner sites, starting with Airtable, GitLab, HubSpot, Notion, Supabase, and Vercel.
When you connect a supported plugin from the ChatGPT plugin directory, you can use Sign in with ChatGPT to create or link an account with that service in fewer steps. On participating partner sites, you can also select Sign in with ChatGPT to create or access your account.
This makes it easier to start working with supported tools in ChatGPT and Codex.
When you sign in, the partner receives only your name, email address, and profile picture, if available. You must still review and approve each plugin’s requested access as a separate step.
-
Codex CLI 0.146.0
$ npm install -g @openai/codex@0.146.0View details
New Features
- Name new sessions with
/newor/clear, pin important threads, and switch between side conversations without closing them. (#34605, #34840, #35011) - Support Agent Plugins manifests, workspace plugin publishing, and additional plugin marketplaces for Amazon Bedrock and Claude Code. (#35105, #35254, #34931, #34979)
- Fork threads with paginated history, including temporary forks that do not appear in thread listings. (#35220, #35251)
- Connect app-server to remote Code Mode hosts over WebSocket. (#35078, #35098)
- Enable standalone web search for compatible custom model providers. (#34846)
- Discover executor-provided skills and securely read their associated resources, including explicitly selected skills. (#35184, #35198)
Bug Fixes
- Honor configured proxies across authentication, plugin downloads, MCP authorization, remote execution, WebSockets, redirects, and LM Studio connections. (#34479, #34509, #34655, #34678, #35023, #35056, #35239)
- Keep MCP connections and Apps tools current when authentication or configuration changes, reconnecting closed servers without restarting healthy connections. (#34952, #34957, #35028, #35144, #35146, #35151)
- Preserve submitted messages, final responses, failed-turn errors, imported timestamps, and approval settings across interruptions, replay, imports, and forks. (#34839, #34777, #35524, #34989, #34664)
- Improve terminal responsiveness and rendering, including nonblocking interrupts, keyboard handling, narrow layouts, hyperlinks, and refreshed mention results. (#35000, #35021, #34775, #34778, #35365, #35375)
- Fix Windows navigation keys, reliably terminate sandboxed process trees, and preserve proxy settings during security reviews. (#34625, #34624, #35036)
- Retain more available skills under tight context budgets and warn when skill catalogs must be truncated. (#34732, #34738, #34997)
Documentation
- Document shared HTTP-client usage, proxy-aware connection pooling, and safe outbound request handling. (#34669)
- Clarify Windows drive-letter canonicalization for
PathUrivalues. (#34667)
Chores
- Publish release artifacts, channel metadata, and installer aliases through OpenAI-hosted release infrastructure, with GitHub fallback. (#34505, #34508, #34729, #34910)
- Sign and notarize bundled macOS helper executables before packaging. (#35264)
- Reduce app-server serialization overhead and unnecessary request-building allocations. (#34761, #34766, #34825)
- Add enterprise-plan recognition and administrator controls for in-app updates. (#35238, #35537)
Changelog
Full Changelog: rust-v0.145.0...rust-v0.146.0
- #34447 Add a route-aware HTTP client pool @copyberry
- #34449 Make external session detection limits configurable @copyberry
- #34451 Attribute external agent imports by provider @copyberry
- #34463 Support alpha hotfix release versions @copyberry
- #34469 Preserve thread settings for goal-first and forked threads @copyberry
- #34476 Separate HTTP execution from request logging @copyberry
- #34478 Honor
CARGO_HTTP_CAINFOin managed proxy environments @copyberry - #34479 Re-resolve system proxy routes across redirects @copyberry
- #34481 Add route-aware redirect test coverage @copyberry
- #34483 Expand route-aware proxy redirect coverage @copyberry
- #34490 Route backend requests through the HTTP client factory @copyberry
- #34491 Route cloud environment discovery through the HTTP client pool @copyberry
- #34495 Honor system proxy settings in the daemon updater @copyberry
- #34497 Preserve custom arg0 for sandboxed exec-server processes @copyberry
- #34505 Mirror Rust release artifacts to Cloudflare R2 @copyberry
- #34506 Respect system proxies during plugin startup sync @copyberry
- #34508 Publish release metadata to R2 channels @copyberry
- #34509 Honor system proxy settings for remote plugins @copyberry
- #34514 Add an optional releases.openai.com installer source @copyberry
- #34516 Allow
numerin codespell checks @copyberry - #34517 Pass empty inherited FDs in the Wine PTY test @copyberry
- #34522 Split MCP connection manager into focused modules @copyberry
- #34525 Add step-scoped data to extension contributors @copyberry
- #34533 Centralize compacted rollout item construction @copyberry
- #34540 Detach Git metadata commands from stdin @copyberry
- #34544 Size Noise handshake buffers to their messages @copyberry
- #34547 Add reciprocal rank fusion skill selection @copyberry
- #34550 Test thread-scoped MCP refresh behavior @copyberry
- #34551 Simplify TUI restoration for the external editor @copyberry
- #34552 Remove unused RtOptions setters @copyberry
- #34553 Remove the unused TUI shutdown app command @copyberry
- #34558 Remove obsolete ignored tests @copyberry
- #34559 Add backend client support for Codex user settings @copyberry
- #34561 Extract MCP binding clients from the connection manager @copyberry
- #34562 Record rollout boundaries for materialized turns @copyberry
- #34563 Page through inherited thread history @copyberry
- #34566 Protect fork history references during rollout cleanup @copyberry
- #34570 Highlight CUDA files as C++ in the TUI @copyberry
- #34573 Accept
forceRefetchin plugin list requests @copyberry - #34578 Gate the TUI suspend restore helper on Unix @copyberry
- #34581 Add routing-card lexical skill selection @copyberry
- #34588 Bind MCP calls to captured catalog revisions @copyberry
- #34590 Add keyed shell environment policy filters @copyberry
- #34597 Enforce exact values from managed config requirements @copyberry
- #34598 Skip missing paths in filesystem sandbox entries @copyberry
- #34601 Sanitize skill names in injection metrics @copyberry
- #34603 Allow explicitly permitted loopback proxy targets @copyberry
- #34605 Allow naming sessions with
/newand/clear@copyberry - #34611 Add compatibility policies for skill catalog rendering @copyberry
- #34612 Detach non-interactive subprocesses from stdin @copyberry
- #34613 Route Windows sandbox proxy traffic by restricting SID @copyberry
- #34615 Initialize missing-path behavior in exec-server sandbox test @copyberry
- #34620 Add exec-server network policy callback types @copyberry
- #34621 Load paginated model context across rollout lineages @copyberry
- #34622 Increase the auto-review model override test timeout @copyberry
- #34624 Terminate Windows process trees with job objects @copyberry
- #34625 Fix Windows TUI navigation key handling @copyberry
- #34626 Scale skill metadata budgets with model context windows @copyberry
- #34629 Harden Windows elevated sandbox startup @copyberry
- #34630 Add a policy-aware HTTP client builder @copyberry
- #34631 Migrate agent identity to the shared HTTP client @copyberry
- #34636 Keep the TUI open when starting a turn fails @copyberry
- #34637 Attribute review findings to repository rules @copyberry
- #34640 Update Windows process-tree tests for inherited FDs @copyberry
- #34641 Harden managed proxy setup for sandboxed executions @copyberry
- #34643 Migrate login HTTP construction to
HttpClient@copyberry - #34644 Verify Git plugin SHA checkouts @copyberry
- #34645 Always assign response item IDs @copyberry
- #34649 Propagate resolved proxy policy through auth routing @copyberry
- #34650 Require auth managers to receive routing configuration @copyberry
- #34651 Migrate core test support to the shared HTTP client @copyberry
- #34654 Render turn diffs for foreign environment paths @copyberry
- #34655 Honor configured proxy routes for auth refreshes @copyberry
- #34664 Preserve approvals reviewer when forking threads @copyberry
- #34667 Document
PathUridrive letter canonicalization @copyberry - #34669 Expand codex-http-client usage guidance @copyberry
- #34678 Route LM Studio requests through the shared HTTP client @copyberry
- #34681 Add session headers to realtime conversation starts @copyberry
- #34687 Configure Codex Auto Review model metadata @copyberry
- #34708 Rename the MCP connection manager to
McpConnectionSet@copyberry - #34713 Order unified exec lifecycle events reliably @copyberry
- #34728 Skip Git enrichment for prewarm and Guardian turns @copyberry
- #34729 Publish stable installer aliases to R2 @copyberry
- #34732 Preserve skill catalog entries under metadata pressure @copyberry
- #34733 Make MCP resource clients follow the latest runtime @copyberry
- #34734 Remove step-scoped data from extension contributors @copyberry
- #34738 Drop skill descriptions before omitting catalog entries @copyberry
- #34744 Update skills budget tests for extension API changes @copyberry
- #34746 Match core skill ordering in extension catalogs @copyberry
- #34747 Register the MCP 2026-07-28 feature flag @copyberry
- #34761 Reduce app-server JSON serialization overhead @copyberry
- #34763 Retry websocket requests when the previous response is missing @copyberry
- #34766 Reduce typed app-server request serialization overhead @copyberry
- #34769 Add the git attribution extension @copyberry
- #34770 Enable exec-server network policy callbacks @copyberry
- #34771 Size unified mention popups to visible results @copyberry
- #34772 Normalize whitespace-only lines in agent messages @copyberry
- #34775 Clamp session headers to narrow terminal widths @copyberry
- #34777 Include the final agent message in turn completion summaries @copyberry
- #34778 Coalesce wrapped OSC 8 hyperlinks in the TUI terminal @copyberry
- #34779 Use the live parent history mode when forking agents @copyberry
- #34781 Upgrade Bazel Rust and LLVM dependencies @copyberry
- #34784 Reject dynamic environments named
local@copyberry - #34785 Report skill catalog truncation during rendering @copyberry
- #34786 Simplify app-server integration test setup @copyberry
- #34789 Avoid unnecessary post-sampling token estimates @copyberry
- #34795 Remove obsolete step store from git attribution tests @copyberry
- #34796 Skip syntax highlighting for lines over 4 KiB @copyberry
- #34797 Suppress omission notices in core-compatible skill catalogs @copyberry
- #34806 Use path URIs in shell approval keys @copyberry
- #34808 Centralize SQLite connection configuration @copyberry
- #34811 Fix network access rendering in sandbox prompts @copyberry
- #34814 Consolidate thread startup around
StartThreadOptions@copyberry - #34816 Support configurable realtime BEM channel prefixes @copyberry
- #34819 Enable git attribution across Codex entry points @copyberry
- #34823 Run code-mode tests in non-Windows Bazel CI @copyberry
- #34824 Normalize Guardian review cwd reuse keys @copyberry
- #34825 Reduce cloning when building Responses requests @copyberry
- #34827 Remove Windows Bazel lint toolchain overrides @copyberry
- #34831 Flush analytics before in-process app server shutdown @copyberry
- #34835 Track compaction time in turn profiles @copyberry
- #34839 Preserve user input when MCP startup is interrupted @copyberry
- #34840 Add persisted thread pinning to the app server @copyberry
- #34844 Remove first-party type from app metadata @copyberry
- #34845 Track multi-agent mode in world state @copyberry
- #34846 Allow custom providers to opt into standalone web search @copyberry
- #34847 Use Guardian model limits for review sessions @copyberry
- #34849 Cache remote plugin catalogs by scope @copyberry
- #34850 Disable image generation for Free-plan accounts @copyberry
- #34851 Use batch metadata for plugin app summaries @copyberry
- #34852 Wake sleeping threads for queued agent mail @copyberry
- #34877 Wait for local plugin cache refreshes in
plugin/list@copyberry - #34883 Set a default user agent for MCP HTTP requests @copyberry
- #34887 Allow disabling the multi-agent wait tool @copyberry
- #34910 Prefer releases.openai.com in standalone installers @copyberry
- #34930 Centralize thread MCP state in
McpRuntime@copyberry - #34931 Use the API plugin marketplace for Amazon Bedrock @copyberry
- #34940 Keep session defaults static during config batch writes @copyberry
- #34952 Reuse MCP connections across runtime refreshes @copyberry
- #34957 Replace closed MCP connections during reconciliation @copyberry
- #34959 Handle @ in local marketplace paths @copyberry
- #34962 Move MCP connection helpers into the test module @copyberry
- #34969 Keep the sleep tool outside code mode @copyberry
- #34978 Honor disabled redirects in route-aware HTTP clients @copyberry
- #34979 Infer the bundled Claude Code plugin marketplace @copyberry
- #34981 Record externally completed agent config imports @copyberry
- #34986 Enforce single-writer ownership for paginated threads @copyberry
- #34989 Preserve timestamps when importing external agent sessions @copyberry
- #34991 Allow omitting MCP tool prefixes per server @copyberry
- #34994 Honor the configured SQLite home across state consumers @copyberry
- #34995 Initialize execution environments with the final HTTP policy @copyberry
- #34996 Separate Codex error details from retry metadata @copyberry
- #34997 Warn when skill catalogs exceed their context budget @copyberry
- #35000 Make TUI turn interrupts nonblocking @copyberry
- #35011 Keep side conversations open when switching threads @copyberry
- #35012 Expose remote skill icon URLs through app server @copyberry
- #35013 Support incremental replay of updated thread items @copyberry
- #35015 Align installed app duration metrics with the legacy baseline @copyberry
- #35016 Add trusted plugin script attribution @copyberry
- #35020 Attribute command executions to trusted plugin scripts @copyberry
- #35021 Adapt keyboard event reporting to the terminal @copyberry
- #35023 Route exec-server HTTP through configured proxy policy @copyberry
- #35028 Preserve refreshed Apps tools across MCP runtime updates @copyberry
- #35029 Preserve plugin attribution across command approvals @copyberry
- #35031 Enforce writer ownership for thread archive and deletion @copyberry
- #35033 Expose Browser Use requirements through the app server @copyberry
- #35034 Route environment registry requests through the shared HTTP client @copyberry
- #35036 Preserve Windows sandbox proxy settings in guardian sessions @copyberry
- #35048 Track app/read request duration @copyberry
- #35049 Register the Guardian V2 feature flag @copyberry
- #35054 Allow disabling the update_plan tool @copyberry
- #35056 Route exec-server WebSockets through configured proxies @copyberry
- #35059 Decouple exec-server HTTP from reqwest types @copyberry
- #35063 Track deferred tool namespaces in world state @copyberry
- #35065 Avoid duplicating deferred sources in tool search @copyberry
- #35067 Fix Bazel test configuration for platform-specific data @copyberry
- #35078 Add WebSocket transport to the code-mode host @copyberry
- #35098 Support remote code-mode hosts in app-server @copyberry
- #35105 Support Agent Plugins manifests @copyberry
- #35106 Allow hosts to customize
wait_for_environmentdescriptions @copyberry - #35144 Prewarm MCP runtime updates in the background @copyberry
- #35146 Refresh MCP runtimes when session auth changes @copyberry
- #35151 Reconnect MCP servers on explicit refresh @copyberry
- #35164 Encapsulate MCP refresh coordination @copyberry
- #35168 Route extension warnings to app-server threads @copyberry
- #35172 Compact host skill paths under metadata pressure @copyberry
- #35175 Wait for reloaded worker completion in the resume test @copyberry
- #35184 Expose executor skills through skill tools @copyberry
- #35194 Preserve output from hooks that exit before reading stdin @copyberry
- #35196 Make the Apps recovery exposure test deterministic @copyberry
- #35198 Enable resource reads for explicit executor skills @copyberry
- #35204 Refresh MCP runtimes across thread startup @copyberry
- #35205 Use current MCP authority for elicitation reviews @copyberry
- #35213 Refresh managed MCP requirements for active threads @copyberry
- #35216 Refresh MCP config independently across threads @copyberry
- #35220 Support paginated thread forks @copyberry
- #35221 Avoid persisting non-local threads for hook transcripts @copyberry
- #35238 Support the ent26 enterprise plan @copyberry
- #35239 Route MCP auth discovery through runtime HTTP clients @copyberry
- #35251 Support ephemeral forks of paginated threads @copyberry
- #35254 Expose workspace plugin publish capability @copyberry
- #35261 Propagate remote plugin IDs to skill metadata @copyberry
- #35262 Track remote plugin IDs in skill invocation analytics @copyberry
- #35264 Sign bundled macOS helper binaries @copyberry
- #35266 Allow disabling the in-process code-mode host fallback @copyberry
- #35267 Harden network approval cancellation and concurrency @copyberry
- #35271 Include code-mode tool names in Responses Lite metadata @copyberry
- #35275 Trace remote exec-server connection setup @copyberry
- #35280 Skip plugin MCP filtering when no allowlists are configured @copyberry
- #35359 Handle exec-server network policy requests in the client @copyberry
- #35363 Include item start times in completion events @copyberry
- #35364 Bound Code Mode metadata compatibility headers @copyberry
- #35365 Keep unified mention results fresh @copyberry
- #35375 Make the keymap action menu responsive @copyberry
- #35408 Ignore generated system skills in the skills watcher @copyberry
- #35414 Raise the MCP server recursion limit @copyberry
- #35523 Shut down the in-process outbound router explicitly @copyberry
- #35524 Preserve terminal turn errors in replayed history @copyberry
- #35525 Skip inactive TUI threads without pending user interaction @copyberry
- #35530 Track model and personality in world state @copyberry
- #35537 Add managed policy for in-app updates @copyberry
- Name new sessions with
-
ChatGPT for iOS 1.2026.202
Improvements and bug fixes
- Voice conversations now use your selected ChatGPT voice and show usage-limit warnings.
- Improved task reconnection and continuity when returning to the app or unlocking with Face ID.
- Composer autocomplete now matches desktop plugin mentions and includes skills from installed plugins.
- Selected-text references remain available after sending so you can preview them again.
- Improved goal controls with clearer progress when pausing or resuming.
- Inline visualizations now render tables and visual themes more reliably.
- Large workspace diffs are more responsive.
- Fixed restored tasks changing the selected model.
- Prevented the composer from becoming stuck after a prompt started.
- Corrected browser and computer tool labels, icons, and placeholder output.
-
ChatGPT Voice and multi-folder projects 26.715
Powered by GPT-Live, ChatGPT Voice lets you talk through work and coordinate tasks in Chat, Work, and Codex in the ChatGPT desktop app.
Start a new chat or task in voice mode, then ask ChatGPT to start, check, or steer work in other threads. On macOS, turn on Screen context to share an appshot of your frontmost window.
Voice is available with Plus, Pro, Business, Edu, and Enterprise plans in the desktop app and through Remote on iOS.
Local projects in the ChatGPT desktop app can now include multiple related folders. From a project’s menu, select Edit project to add folders and choose the primary folder. New chats, Git operations, and automatic discovery of
AGENTS.md, skills, andconfig.tomluse the primary folder. Secondary folders remain available for file search, reading, and editing.Get started with ChatGPT Voice and multi-folder local projects.
-
Codex CLI 0.145.0
$ npm install -g @openai/codex@0.145.0View details
New Features
- Added experimental paginated thread history with efficient resume, search, persisted names, sub-agent support, and memories. (#33364, #33907, #34085, #34229, #34386)
- Expanded
/importto migrate Cursor and Claude Code settings, MCP servers, plugins, sessions, commands, and project-scoped memories. (#31672, #33411, #33426, #33444) - Added experimental Amazon Bedrock login, custom endpoint and authentication support, and GPT-5.6 Sol as the default Bedrock model. (#31327, #33170, #33175, #32288, #33695)
- Added audio inputs and tool outputs, including common local audio formats, and introduced streaming realtime V3 conversations. (#33261, #33856, #33932, #34080, #34385)
- Stabilized the opt-in multi-agent V2 experience with configurable sub-agent models, reasoning levels, concurrency, restored roles, and improved agent navigation. (#33550, #33631, #33657, #33841, #34383)
- Added secure, clickable inline visualization links in the terminal UI. (#33925, #34217, #34346)
Bug Fixes
- Editing an earlier prompt or retrying a safety-buffered turn now creates a contextual branch, preserving the original conversation, attachments, and mention bindings. (#33201, #33207, #33211)
- Improved terminal responsiveness for long conversations and streamed output through incremental Markdown rendering, fewer redraws, caching, and bounded command output. (#34045, #34049, #34216, #34223, #34359)
- Prevented slow or conflicting MCP startup and authentication flows by enforcing startup timeouts, avoiding blocking OAuth discovery, serializing refreshes, and reusing tool catalogs safely. (#32229, #32781, #32825, #33184, #33297)
- Improved Windows execution and sandbox reliability, including native exec-server sandboxing, network-proxy enforcement, hidden helper consoles, and correctly quoted hook commands. (#32849, #32857, #33926, #34423)
- Fixed compact release-metadata parsing and macOS code-mode installation, with an in-process fallback when the external code-mode host is unavailable. (#31667, #31876, #31899)
- Strengthened safety and approval handling with better forced-
rmdetection, consistent full-access confirmation, and preserved rejection reasons across tools. (#32989, #33464, #34400)
Documentation
- Updated the bundled OpenAI Docs skill with current GPT-5.6 model resolution, prompting, and migration guidance across macOS, Linux, and Windows. (#31842, #33121)
Chores
- Migrated bundled GPT-5.4 selections and internal uses to the corresponding GPT-5.6 Terra and Luna variants. (#33173)
- Reduced startup and large-context overhead with concurrent skill/plugin discovery and more efficient remote compaction. (#31566, #33369, #33423, #34431)
- Updated the packaged ripgrep binary to 15.2.0. (#34384)
Changelog
Full Changelog: rust-v0.144.0...rust-v0.145.0
- #31667 fix: parse compact release metadata in installer @efrazer-oai
- #31362 core: route realtime and memories through HTTP client factory @bolinfest
- #31566 perf(skills): reuse walk inventory for host loading @jif-oai
- #31576 Bound exec-server process event reordering @jif-oai
- #31756 test(skills): assert symlinked metadata loading @jif-oai
- #31581 Resolve selected capability roots without starting executors @jif-oai
- #31789 Stop persisting RMCP service traces @jif-oai
- #31792 Summarize streamed response item logs @jif-oai
- #31791 Filter routine Hyper logs from SQLite @jif-oai
- #31790 Reduce MCP tool-list trace volume @jif-oai
- #31804 Stabilize the memories feature flag @jif-oai
- #31803 fix(mcp): default Apps product SKU to codex @alecbarber-oai
- #31745 code-mode: retain shared MCP types for deferred tools @sayan-oai
- #31672 Import enabled plugins from known marketplaces @charlesgong-openai
- #31652 fix(tui): hide empty reasoning summaries @fcoury-oai
- #31767 Remove the network proxy config wrapper @jif-oai
- #31481 fix: forward originator to Codex Apps MCP @raquel-openai
- #31363 codex-api: route file uploads through HTTP client factory @bolinfest
- #31813 tui: update safety buffering copy @etraut-openai
- #31830 fix(sandboxing): initialize network proxy config inline @fcoury-oai
- #31431 build: ratchet direct reqwest dependencies @bolinfest
- #31876 code-mode: fix installation on darwin @cconger
- #31842 Update bundled OpenAI Docs skill for GPT-5.6 @kkahadze-oai
- #31637 login: route raw auth flows through HTTP client @bolinfest
- #31686 [codex-apps] Filter optional file fields by tool schema @tsarlandie-oai
- #31899 code-mode: fall back to using in process v8 if we fail to resolve external process @cconger
- #31805 Bound remote MCP stdio lines @jif-oai
- #30293 Resolve and pin MCP OAuth credential stores @stevenlee-oai
- #31892 exec-server: materialize filesystem workspace roots @pakrym-oai
- #31327 feat: add managed Bedrock login API @celia-oai
- #31295 bench: add codex help e2e macrobenchmark @anp-oai
- #31428 bench: add e2e benchmark entrypoints @anp-oai
- #31937 exec-server: expose process helper to outer sandbox @pakrym-oai
- #32093 Remove the legacy exec policy engine @copyberry
- #32106 Reduce startup latency for ancestor discovery @copyberry
- #32112 Bound streamed exec-server HTTP response bodies @copyberry
- #32122 Test the shared exec-server HTTP response byte budget @copyberry
- #32123 Bound exec-server stdio JSON-RPC messages @copyberry
- #32126 Test bounded concurrency in ancestor discovery @copyberry
- #32134 Test stdio JSON-RPC size limits with LF and CRLF @copyberry
- #32135 Propagate tracing subscribers to exec start tasks @copyberry
- #32150 Keep unified exec output collection bounded @copyberry
- #32193 Validate memory consolidation artifacts before succeeding @copyberry
- #32197 Rebind memory consolidation workspace roots @copyberry
- #32200 Add a skill invocation extension contributor @copyberry
- #32206 Always send reasoning parameters in Responses requests @copyberry
- #32213 Generate unique IDs for review rollout messages @copyberry
- #32214 Propagate workspace roots to exec-server sandboxes @copyberry
- #32229 Serialize MCP OAuth credential refreshes @copyberry
- #32231 Support pending remote environment registration @copyberry
- #32232 Let permission hooks resolve strict auto-review requests @copyberry
- #32234 Add dedicated storage for paginated thread history @copyberry
- #32246 Extract reverse JSONL scanning from session indexing @copyberry
- #32256 Improve Responses WebSocket timing telemetry @copyberry
- #32261 Preserve local path conventions in automatic approvals @copyberry
- #32263 Include start times in terminal turn events @copyberry
- #32272 Expose scheduled tasks in plugin details @copyberry
- #32274 Remove the personality migration @copyberry
- #32276 Repair unterminated rollout files before appending @copyberry
- #32277 Honor
personality = "none"in model instructions @copyberry - #32280 Include terminal errors in turn completion events @copyberry
- #32286 Clarify waiting behavior in safety buffering prompts @copyberry
- #32288 Make GPT-5.6 Sol the default Bedrock model @copyberry
- #32289 Persist paginated items in the local thread store @copyberry
- #32290 Respect model support for reasoning summaries @copyberry
- #32301 Trust hooks from materialized workspace plugins @copyberry
- #32302 Prefer the Codex home socket for Unix IDE context @copyberry
- #32305 Improve file blob upload diagnostics @copyberry
- #32312 Require prefixes for outbound response item IDs @copyberry
- #32316 Stop falling back to older model availability announcements @copyberry
- #32326 Use canonical links in the moved config notice @copyberry
- #32332 Add ordinals to paginated rollout records @copyberry
- #32441 Preserve parent sandbox enforcement for memory consolidation @copyberry
- #32460 Emit thread-idle lifecycle after guardian interrupts @copyberry
- #32461 Expand tabs when rendering TUI diffs @copyberry
- #32485 Use available width for skill names in the toggle view @copyberry
- #32628 Improve composer completion target resolution @copyberry
- #32698 Extract connector runtime snapshot management @copyberry
- #32744 Log missing personality messages at trace level @copyberry
- #32746 Make advanced reasoning selection explicit in the TUI @copyberry
- #32747 Align Guardian reviews with session configuration @copyberry
- #32749 Expose model overrides for multi-agent v2 spawns @copyberry
- #32751 Restrict spawned-agent models to the active backend @copyberry
- #32761 Add shadow metrics for lexical skill selection @copyberry
- #32768 Align shadow skill selection with observable sources @copyberry
- #32780 Enable skill search shadow selection by default @copyberry
- #32781 Apply MCP startup timeouts during client creation @copyberry
- #32801 Refactor OAuth store lock contention tests @copyberry
- #32822 Make explicit multi-agent mode override proactive delegation @copyberry
- #32825 Avoid blocking thread startup on MCP OAuth discovery @copyberry
- #32835 Forward turn metadata in standalone web search @copyberry
- #32837 Restore V2 agent identities on root thread resume @copyberry
- #32838 Reap exited PID-managed app-server children @copyberry
- #32844 Expand millisecond duration histogram boundaries @copyberry
- #32849 Hide Windows filesystem helper console windows @copyberry
- #32857 Require the elevated Windows sandbox for network proxies @copyberry
- #32858 Persist slash-command popup dismissal @copyberry
- #32864 Coalesce concurrent Windows sandbox setup requests @copyberry
- #32866 Allow responses after image generation @copyberry
- #32867 Include connector IDs in MCP tool call analytics @copyberry
- #32875 Use model catalog policies for Guardian auto review @copyberry
- #32881 Broaden remote compaction model fallback @copyberry
- #32884 Prepare external agent migration for source adapters @copyberry
- #32887 Tag shell tool telemetry by command category @copyberry
- #32891 Attach connector caches to diagnostic uploads @copyberry
- #32894 Serialize plugin install requests @copyberry
- #32896 Load model context from a bounded rollout suffix @copyberry
- #32897 Route blocked network requests to their owning calls @copyberry
- #32898 Expose structured standalone web search results @copyberry
- #32899 Add exec-server environment status checks @copyberry
- #32900 Derive collaboration settings from turn context @copyberry
- #32903 Include session IDs in tool item analytics events @copyberry
- #32905 Timestamp app-server notifications at emission @copyberry
- #32911 Allow injecting the models manager into
ThreadManager@copyberry - #32920 Expose environment status through app-server @copyberry
- #32923 Materialize paginated thread history in SQLite @copyberry
- #32928 Resume thread history projection from its SQLite checkpoint @copyberry
- #32945 Restrict Guardian reviewer tools @copyberry
- #32949 Tighten recommended plugin install suggestions @copyberry
- #32952 Scope runtime workspace roots to execution environments @copyberry
- #32985 Expose exact per-response usage in raw app-server events @copyberry
- #32989 Always confirm before enabling full access @copyberry
- #33013 Bound exec-server JSON-RPC decoding complexity @copyberry
- #33026 Include raw response completions in TypeScript envelopes @copyberry
- #33030 Remove task messages from
list_agentsoutput @copyberry - #33031 Preserve JSON number precision in exec-server RPC messages @copyberry
- #33035 Use session IDs for prompt cache keys @copyberry
- #33040 Send plugin analytics with API key authentication @copyberry
- #33076 Add an agent extension runner @copyberry
- #33093 Preserve streamed output during capped history replay @copyberry
- #33105 Fix TUI status visibility around streamed output @copyberry
- #33107 Preserve special filesystem subpaths as wire strings @copyberry
- #33109 Reject forks of paginated threads @copyberry
- #33113 Allow injecting the Codex Apps tools cache @copyberry
- #33121 Refine GPT-5.6 prompting and migration guidance @copyberry
- #33147 Support model catalog permission messages @copyberry
- #33149 Build MCP tool runtimes before router planning @copyberry
- #33150 Clarify exec yield timing on Windows @copyberry
- #33152 Support paginated thread history in app-server list APIs @copyberry
- #33155 Trace startup prewarm tasks @copyberry
- #33156 Run detached reviews as review-agent turns @copyberry
- #33159 Move sleep items to the extension-owned lifecycle path @copyberry
- #33166 Defer Noise environment connections until registration @copyberry
- #33167 Document the Windows exec yield time range @copyberry
- #33170 Support Amazon Bedrock login in the app server @copyberry
- #33173 Migrate GPT-5.4 uses to GPT-5.6 variants @copyberry
- #33175 Handle Amazon Bedrock credentials during logout @copyberry
- #33177 Support model catalog templates for Guardian policy prompts @copyberry
- #33180 Serialize concurrent MCP stdin writes @copyberry
- #33182 Preserve plugin install failure subtypes during imports @copyberry
- #33184 Reuse MCP tool catalogs across sessions @copyberry
- #33185 Keep approval test targets in the temporary home @copyberry
- #33187 Honor workspace spend controls in rate-limit handling @copyberry
- #33198 Keep interrupted prompts in conversation history @copyberry
- #33200 Separate exec permission paths from core models @copyberry
- #33201 Branch conversations when editing earlier TUI prompts @copyberry
- #33203 Preserve in-flight state when restoring thread input @copyberry
- #33207 Retry safety-buffered turns on a forked thread @copyberry
- #33209 Separate session state from session I/O @copyberry
- #33211 Preserve thread context when retrying or editing turns @copyberry
- #33213 Prepare Python SDK 0.144.4 stable release @copyberry
- #33223 Instrument environment and plugin resolution paths @copyberry
- #33232 Disambiguate skill mentions from shell parameters @copyberry
- #33237 Fix skill completion around bound mentions with suffixes @copyberry
- #33239 Render TUI composer tabs as single-column spaces @copyberry
- #33243 Add auto-compaction fallback token-budget settings @copyberry
- #33251 Report selected environment connection transitions @copyberry
- #33255 Add a fallback phase before automatic context rollover @copyberry
- #33261 Add Frameless Bidi support for realtime conversations @copyberry
- #33297 Allow MCP servers to opt out of tool catalog caching @copyberry
- #33308 Expand MCP tool catalog cache regression coverage @copyberry
- #33364 Enable paginated thread history in app-server @copyberry
- #33367 Respect final-answer boundaries for queued agent mail @copyberry
- #33369 Scan skill roots concurrently @copyberry
- #33373 Render TUI prompts before submitting user turns @copyberry
- #33411 Migrate plugin commands into skills on install @copyberry
- #33412 Refactor world-state rendering tests into snapshots @copyberry
- #33414 Expose connector candidates from imported sessions @copyberry
- #33421 Fetch workspace connectors concurrently @copyberry
- #33423 Load executor plugin declarations concurrently @copyberry
- #33424 Attribute OpenAI docs MCP requests to Codex @copyberry
- #33425 Refresh host skill catalogs through world state @copyberry
- #33426 Add Cursor support to setup import @copyberry
- #33427 Propagate deferred environment capability roots to MCP @copyberry
- #33430 Avoid creating metadata paths in the Windows sandbox @copyberry
- #33432 Preserve paginated history for spawned subagents @copyberry
- #33435 Warn on conflicting capability root locations @copyberry
- #33441 Shut down Codex threads after approval scenarios @copyberry
- #33444 Add external agent memory migration @copyberry
- #33445 Select the elevated Windows sandbox for network proxies @copyberry
- #33446 Remove the unused network proxy loader @copyberry
- #33454 Track prompt cache write token usage @copyberry
- #33456 Move external agent migration into its crate @copyberry
- #33457 Use final answers in turn history summaries @copyberry
- #33459 Allow more time for image generation in code mode @copyberry
- #33464 Strengthen forced
rmcommand detection @copyberry - #33467 Remove template IDs from MCP tool call metadata @copyberry
- #33500 Add cache-write tokens to the raw response schema @copyberry
- #33509 Preserve encrypted content in MCP tool outputs @copyberry
- #33550 Unify multi-agent settings under
agents@copyberry - #33572 Expose spawn agent types only when roles are configured @copyberry
- #33605 Add fielded BM25 to shadow skill selection @copyberry
- #33613 Add character n-gram skill selection @copyberry
- #33614 Add multi-query lexical skill selection @copyberry
- #33631 Honor configured model defaults for spawned agents @copyberry
- #33632 Remove generated-default filesystem path variants @copyberry
- #33633 Clarify when to wait for starting environments @copyberry
- #33636 Clarify when to wait for starting environments @copyberry
- #33639 Remove the unused realtime WebRTC crate @copyberry
- #33640 Avoid duplicate cached app list update notifications @copyberry
- #33645 Run
write_stdinconcurrently across terminal sessions @copyberry - #33651 Add an app-server API for reading app metadata @copyberry
- #33656 Validate reasoning effort after applying spawn roles @copyberry
- #33657 Restore agent roles when reloading v2 sub-agents @copyberry
- #33658 Keep active-turn environments stable across settings updates @copyberry
- #33659 Require data URLs for code-mode image output @copyberry
- #33665 Refresh step world state for all sessions @copyberry
- #33677 Forward thread originators from standalone extensions @copyberry
- #33680 Reword the apply_patch tool description @copyberry
- #33683 Preserve scope and provenance for imported agent memory @copyberry
- #33684 Extract TUI approval request payloads into structs @copyberry
- #33687 Avoid unnecessary writes during migration repair @copyberry
- #33695 Support custom transports for Amazon Bedrock @copyberry
- #33841 Make parent-owned sub-agent threads read-only in the TUI @copyberry
- #33842 Give the zsh fork decline test more execution time @copyberry
- #33843 Add an API for reading installed app runtime state @copyberry
- #33845 Confirm usage-limit resets before redemption @copyberry
- #33848 Fix the managed Bedrock logout test assertion @copyberry
- #33851 Record web search result payload sizes @copyberry
- #33852 Add batched executor capability discovery @copyberry
- #33855 Tag realtime transcript tail flush delegations @copyberry
- #33856 Stream realtime V3 Codex handoff output @copyberry
- #33858 Isolate core tests from shell and rollout persistence @copyberry
- #33861 Test workspace write isolation across exec servers @copyberry
- #33862 Suppress empty multi-agent mode messages @copyberry
- #33863 Report detailed session import error types @copyberry
- #33864 Keep feature tests focused on behavior @copyberry
- #33866 Remove the redundant tool dispatch wrapper @copyberry
- #33867 Add grace period to code-mode yield timeouts @copyberry
- #33868 Remove stale ignored core tests @copyberry
- #33870 Remove the redundant borrowed line wrapping helper @copyberry
- #33872 Remove unused TUI collaboration mode indicators @copyberry
- #33876 Track collaboration mode instructions in world state @copyberry
- #33883 Report CLI as the external agent config import source @copyberry
- #33889 Centralize thread MCP connections in
McpRuntime@copyberry - #33892 Limit rollout metadata reads to headers @copyberry
- #33893 Track realtime conversation state in world state @copyberry
- #33895 Add SessionEnd hooks for thread teardown @copyberry
- #33896 Expose plugin installation interstitial requirements @copyberry
- #33901 Support ChatGPT-branded Desktop app builds @copyberry
- #33902 Add bounded batch lookups for message history @copyberry
- #33903 Route realtime V3 handoffs by response channel @copyberry
- #33905 Batch persistent history reads during reverse search @copyberry
- #33906 Launch managed network proxies on remote executors @copyberry
- #33907 Add occurrence search for paginated threads @copyberry
- #33908 Allow publishing plugins through share updates @copyberry
- #33921 Preserve sub-agent liveness in the agent picker @copyberry
- #33922 Allow selecting path-backed agents in the TUI picker @copyberry
- #33923 Add audio variants to user input protocols @copyberry
- #33925 Render inline visualization links in the TUI @copyberry
- #33926 Fix quoted hook commands on Windows @copyberry
- #33929 Handle audio inputs and Bazel unit test arguments @copyberry
- #33930 Track inherited paginated rollout prefixes @copyberry
- #33932 Forward audio inputs to the Responses API @copyberry
- #33938 Centralize SQLite connection configuration @copyberry
- #33944 Track permission instructions in world state @copyberry
- #33950 Let users remember the working directory for resumed sessions @copyberry
- #33961 Refresh bundled model metadata @copyberry
- #33963 Add context to sampling retry logs @copyberry
- #33982 Gate audio history by model input modalities @copyberry
- #34038 Handle compressed rollouts in doctor thread inventory @copyberry
- #34045 Render streamed Markdown incrementally @copyberry
- #34047 Avoid resending the model for reasoning shortcuts @copyberry
- #34049 Avoid redundant TUI redraws while streaming @copyberry
- #34067 Seed realtime V3 sessions with initial text items @copyberry
- #34080 Add audio output support to dynamic tools and code mode @copyberry
- #34085 Support legacy views for paginated thread history @copyberry
- #34194 Avoid cloning thread data when rendering transcripts @copyberry
- #34197 Use the Markdown collector as the streaming source of truth @copyberry
- #34198 Start side conversations without replaying inherited turns @copyberry
- #34199 Avoid liveness races when starting side conversations @copyberry
- #34204 Avoid cloning buffered TUI history lines @copyberry
- #34206 Avoid retaining decoded MCP images in history cells @copyberry
- #34216 Speed up TUI Markdown layout @copyberry
- #34217 Keep incremental rendering with visualization context @copyberry
- #34218 Track TUI command completion separately from output @copyberry
- #34222 Avoid buffering replay-irrelevant thread notifications @copyberry
- #34223 Cache finalized Markdown history rendering @copyberry
- #34224 Avoid cloning file changes in TUI diff rendering @copyberry
- #34226 Backfill completion items only for the active exec turn @copyberry
- #34229 Persist names for paginated threads @copyberry
- #34232 Remeasure dynamic cells in the transcript overlay @copyberry
- #34234 Avoid redundant TUI subagent metadata requests @copyberry
- #34271 Migrate legacy exec policy allow rules @copyberry
- #34293 Preserve zsh tied PATH exports in shell snapshots @copyberry
- #34344 Reject unsupported history modes when loading rollouts @copyberry
- #34345 Remove unused Rust helpers @copyberry
- #34346 Track inline visualization directives during streaming @copyberry
- #34347 Avoid cloning deferred TUI lifecycle payloads @copyberry
- #34348 Cache TUI flex heights across frame passes @copyberry
- #34355 Parallelize TUI bootstrap requests @copyberry
- #34357 Render streamed command output through preview iterators @copyberry
- #34359 Keep streamed command output bounded in the TUI @copyberry
- #34361 Avoid cloning thread history for token usage replay @copyberry
- #34365 Animate Max and Ultra reasoning effort changes @copyberry
- #34366 Avoid cloning hyperlink text during TUI rendering @copyberry
- #34368 Use app-server skill metadata directly in the TUI @copyberry
- #34371 Clear stale Guardian reviews when turns end @copyberry
- #34375 Extend second-based latency histogram buckets @copyberry
- #34378 Avoid rendering generated images twice @copyberry
- #34380 Stop retrying turns with invalid tool images @copyberry
- #34381 Avoid cloning Responses WebSocket payloads @copyberry
- #34382 Keep paginated thread Git metadata in SQLite @copyberry
- #34383 Mark multi-agent v2 as stable @copyberry
- #34384 Update packaged ripgrep to 15.2.0 @copyberry
- #34385 Preserve audio across history and tool outputs @copyberry
- #34386 Enable memories for paginated threads @copyberry
- #34387 Refresh bundled model metadata @copyberry
- #34389 Route Codex Apps MCP through plugin service @copyberry
- #34390 Use copy-on-write storage for history snapshots @copyberry
- #34392 Ignore inherited ACEs when refreshing Windows write roots @copyberry
- #34393 Add configurable hook context spill limits @copyberry
- #34396 Run compact session-start hooks before turn continuation @copyberry
- #34400 Propagate approval rejection reasons @copyberry
- #34403 Update tests for history and hook API changes @copyberry
- #34407 Resolve paginated rollout lineages @copyberry
- #34408 Support threadless MCP connections without event channels @copyberry
- #34409 Limit the Linux
/procpreflight filesystem view @copyberry - #34411 Require absolute paths for test SQLite configuration @copyberry
- #34413 Remove CSV-backed agent jobs @copyberry
- #34416 Show completed hook warnings in TUI headers @copyberry
- #34417 Enrich app/read connector metadata @copyberry
- #34423 Support Windows sandboxing in the exec server @copyberry
- #34429 Move shared skill models into
codex-skills@copyberry - #34431 Optimize remote compaction history handling @copyberry
- #34434 Support catalog messages for non-request approval policies @copyberry
- #34435 Resolve outbound proxy routes explicitly @copyberry
- #34436 Honor managed permission profiles in network proxy resolution @copyberry
- #34438 Increase the patch approval test timeout @copyberry
- #34441 Add buffered code-mode exec yields @copyberry
-
ChatGPT for iOS 1.2026.195
New features
- Added support for rendering Mermaid diagrams inline in task transcripts.
- Added support for interactive forms in Codex tasks.
- Added support for restoring unsent prompts when switching between tasks, hosts, and workspaces.
Improvements and bug fixes
- Improved task lists to sort by recent activity and show unavailable hosts when creating a task.
- Improved the composer with selected-text previews and smoother new-task transitions.
- Improved goals with support for resuming blocked or usage-limited runs.
- Improved plan progress, Fast controls, and inline dictation.
- Improved Remote onboarding, composer guidance, and iPad navigation.
- Fixed an issue that could close the app when duplicate task-list entries appeared while starting a task.
- Fixed iOS 18 task actions and task-list styling.
- Fixed composer spacing, attachment menu padding, and duplicate transcription indicators.
-
Codex CLI 0.144.6
$ npm install -g @openai/codex@0.144.6View details
Bug Fixes
- Refreshed bundled instructions for GPT-5.6 Sol, Terra, and Luna, and corrected their context windows to 272,000 tokens. (#33972, #34009)
Changelog
Full Changelog: rust-v0.144.5...rust-v0.144.6
- #33972 Backport refreshed bundled model metadata to 0.144 @sayan-oai
- #34009 Narrow 0.144 hotfix to GPT-5.6 prompts and context @sayan-oai
-
Codex CLI 0.144.5
$ npm install -g @openai/codex@0.144.5View details
Bug Fixes
- Improved dangerous-command detection, including more forced
rmforms, and provides clearer rejection reasons when commands are denied. (#33455)
Changelog
Full Changelog: rust-v0.144.4...rust-v0.144.5
- #33455 [release/0.144] fix(core) expand is_dangerous_command @dylan-hurd-oai
- Improved dangerous-command detection, including more forced
-
Codex CLI 0.144.4
$ npm install -g @openai/codex@0.144.4View details
Chores
- No user-facing changes in this patch release.
Changelog
Full Changelog: rust-v0.144.3...rust-v0.144.4
-
ChatGPT for iOS 1.2026.188
New features
- Added support for inline visualizations in Codex tasks.
Improvements and bug fixes
- Improved creating and managing tasks from conversations, with reliable links to newly created tasks.
- Improved tool activity styling and progress indicators.
- Improved file-opening feedback.
- Improved the composer so controls remain visible above the keyboard for long prompts and larger text sizes.
- Fixed Fast mode selection and restoration for each task.
- Fixed initial prompts ignoring the selected approval preset.
- Fixed autocomplete backgrounds and task rows becoming unresponsive during swipe gestures.
-
Codex joins the ChatGPT desktop app 26.707
Codex is now part of the ChatGPT desktop app on macOS and Windows. Existing Codex app users can update as usual and keep their projects, settings, and workflows. You can make Codex the default view and, on macOS, keep the Codex app icon.
New features
- Edit Markdown and code directly in the app, use inline annotations, and ask Codex to revise selected content.
- Use PR Chat to review GitHub pull requests and ask Codex about changes in context. Send inline review feedback, inspect proposed patches, and edit, accept, or reject them without leaving the app.
- Connect custom domains to published Sites.
Performance improvements and bug fixes
- Made Computer Use faster with GPT-5.6.
- Made task and subagent activity easier to follow while Codex works.
- Simplified plugin management by moving it into Settings.
- Improved permission handling when resuming tasks or sending follow-ups.
- Added clearer Full access warnings and a dialog when combining Full access with Ultra.
- Improved macOS and Windows setup, including macOS installation, Git-backed workflows, and Computer Use on Windows.
- Fixed task resumption for local projects and onboarding retry loops.
- Fixed scrolling in pull request reviews and expanded Mermaid diagram labels.
- Improved mobile connection reliability and fixed video rendering for SSH projects.
- Additional performance improvements and bug fixes.
-
ChatGPT for iOS 1.2026.181
New features
- Added support for creating, searching, opening, forking, and managing Codex tasks directly from a conversation.
- Added filters for staged, unstaged, branch, and last-turn changes, with controls for comparing branches.
- Added support for adding selected transcript text directly to the composer.
- Added previews for image and file attachments before sending.
- Added inline Photos and Camera pickers to the attachment menu.
- Added a connection shortcut and support for SSH hosts using private keys or no credentials.
- Added usage limits and credit details to the task menu.
Improvements and bug fixes
- Improved the task list with consistent task terminology, clearer delegated task titles, and a Needs input status.
- Improved initial task loading and foreground recovery.
- Improved autocomplete by selecting the first result automatically and accepting it with Return.
- Improved model, reasoning, and Fast settings so changes remain scoped to the current task.
- Improved task-management and dynamic tool activity presentation.
- Improved side chats to open directly when only one conversation is available.
- Improved plugin autocomplete with installed plugins and their icons.
- Improved workspace diff accuracy and expand-and-collapse navigation.
- Improved recovery by preserving thread state across reconnects and host pairings across sign-out.
- Fixed stuck thread-list loading, prompt mode deadlocks, stale images, and microphone permission alerts.
- Fixed shake to undo and keyboard refocusing after sending a prompt.
June 2026
-
Codex Remote reaches general availability
Codex Remote has reached general availability. Use Codex from the ChatGPT mobile app to start or continue work on a connected Mac or Windows host, review progress, and approve actions from your phone.
Remote Control now uses authenticated one-to-one QR pairing between each iOS or Android device and each host. Update the ChatGPT mobile app and Codex App to the latest versions before connecting. Connections used since June 8, 2026, remain paired; older inactive connections need to pair again.
The new DigitalOcean plugin lets Codex provision a DigitalOcean Droplet, configure SSH access, and connect it to the Codex App as a remote workspace.
See Remote connections for setup and troubleshooting.
-
ChatGPT for iOS 1.2026.167
New features
- Added per-host personality settings with Friendly and Pragmatic options.
- Added support for editing goals directly in the composer.
- Added a link from forked conversations back to the original thread.
Improvements and bug fixes
- Improved side chat visibility with separate conversations above the composer.
- Improved composer autocomplete for commands, skills, and plugins from any prefix.
- Improved progress visibility for subagents, tasks, and worktree creation.
- Fixed long threads loading.
- Improved workspace file search, code review drafts, steering, and host setup and recovery.
- Fixed Face ID unlocking, stopping responses, collapsed sections, and dark-mode host indicators.
-
Record & Replay and remote task handoff 26.616
New features
- Added Record & Replay, a macOS feature that turns a demonstrated workflow into a reusable skill. Initial availability excludes the European Economic Area, the United Kingdom, and Switzerland. You or your administrator must also enable Computer Use.
- Added bulk actions to automation run history so you can mark every run as read or archive eligible runs.
- Added thread handoff between local and remote hosts, so you can move a thread to a matching project on a connected host and continue it there. Codex can also coordinate the handoff for you.
Performance improvements and bug fixes
- Added new deep links to manage SSH connections.
- Improved Browser Use so visible-tab routing and annotations persist when a draft browser session moves to the server.
- Additional performance improvements and bug fixes.
-
More Codex features in the EEA, UK, and Switzerland
More Codex app capabilities are rolling out to users in the European Economic Area, the United Kingdom, and Switzerland:
- Computer Use is available on macOS and Windows in these regions, so Codex can operate desktop apps by seeing, clicking, and typing.
- The Codex Chrome extension is available for browser tasks that need signed-in Chrome context, working across tabs in the background without taking over your browser.
- Memories can remember useful preferences, recurring workflows, tech stacks, and repository conventions when enabled. Memories are off by default in the European Economic Area, the United Kingdom, and Switzerland.
- Chronicle is available as an opt-in research preview for ChatGPT Pro subscribers on macOS, helping Codex build memories from recent screen context.
-
ChatGPT for iOS 1.2026.160
New features
- Added a workspace file browser for previewing files and linking workspace paths into prompts.
- Added a directory picker for choosing a workspace folder when starting a new thread.
- Added controls to expand or collapse all diffs while reviewing changed files.
- Added MCP approval choices for allowing requested actions in the current chat or across chats.
- Added LaTeX rendering in Codex messages and plans.
Improvements and bug fixes
- Improved status indicators for running threads, queued prompts, side chats, and subagents.
- Improved pairing and onboarding with clearer errors, manual pairing-code support, and more reliable host selection after pairing.
- Improved task-list recovery, reconnect state, host-specific refresh, and thread performance.
- Improved Codex profile sharing, activity history, and settings layout.
- Improved goal workflows with a composer shortcut, desktop-aligned goal message actions, and better resumed question handling.
- Improved assistant message actions, transcript layout, and public rate-limit names.
- Fixed stuck thread-list swipe actions, duplicate messages when reopening a new thread, spawned subagents appearing as top-level task rows, and misleading connection errors when sending prompts.
-
Usage resets and Browser Developer mode 26.609
New features
- Added rate-limit reset banking for Plus and Pro users, including one free reset at launch and referral invitations for earning more during the current promotion. Eligible Business members can invite coworkers to earn shared workspace credits through a separate referral program.
- Added Developer mode for Browser use in Chrome and the Codex in-app browser. It gives Codex controlled Chrome DevTools Protocol (CDP) access for performance profiling and deeper debugging of network traffic, console output, runtime errors, and page state.
- Added the
/initcommand to the app composer for creating project instructions with the same initialization workflow as the Codex CLI. - Added customizable macOS Dock icons with light and dark Codex variants.
- Added Computer Use for Enterprise users outside the European Economic Area, the United Kingdom, and Switzerland.
- Added support for configuring per-app access controls for Computer Use on Windows.
- Added an Unread chats section to the command menu, with the most recently updated unread chat selected by default.
Performance improvements and bug fixes
- Made Browser use up to 2x faster through CDP and DOM snapshot optimizations that reduce browser round trips.
- Made command, browser, integration, and source activity summaries easier to understand, and improved how completed chats present files, automations, and other durable output.
- Improved plugin management by including workspace plugins, refreshing plugin state more reliably after installation or removal, and letting you upload a new version of an already-shared plugin without changing its access.
- Improved usage-limit errors with inline plan and workspace guidance, including reset timing when available.
- Added Cmd+Enter and Ctrl+Enter as shortcuts for submitting custom approval feedback.
- Fixed Browser use download handling and improved Developer mode recovery and diagnostics.
- Fixed scheduled automations so they honor the selected approval mode, and fixed manual project ordering, Browser tab dragging, MCP app sizing after right-pane transitions, and clickable ChatGPT thread mentions.
- Fixed issues affecting background agent tab restoration, commit and pull request message generation, sidebar pull request status updates, Codex Mobile QR pairing, remote-control MFA, remote SSH installation and connection, updater prompts, and overlay positioning at non-default zoom levels.
- Additional performance improvements and bug fixes.
-
Codex app 26.608
New features
- Added Import to Codex flows for importing supported setup from Claude Code and Claude Cowork, including during onboarding.
- Revamped plugins screen with separate tabs, marketplace and category filters, keyboard navigation, and clearer install actions.
- Expanded Settings search to find options from more panels, including Git and pets.
Performance improvements and bug fixes
- Fixed goal timer overlap in narrow layouts.
- Reduced unread notifications while an active goal continues running.
- Kept review diff ordering consistent with the file tree.
- Improved window rendering on systems that don’t support translucent backdrops, including Windows 10.
- Additional performance improvements and bug fixes.
-
ChatGPT for iOS 1.2026.153
New features
- Added support for choosing a branch, creating a worktree, and running an environment setup script for new threads.
- Added a Codex profile screen with usage stats and token activity charts.
- Added
/goalsupport for creating and managing goals from Codex Mobile. - Added inline review comments when viewing changed files.
- Added support for asking in side chat from selected transcript text.
- Added support for editing the latest sent prompt.
Improvements and bug fixes
- Improved attachment support on Windows hosts.
- Skills and plugins now appear directly inline in the composer.
- Improved side chat and queued prompt visibility while a thread is running.
- Improved message styling, navigation, tool activity, Face ID behavior, archived-thread browsing, and thread UI polish.
-
Codex app updates 26.602
New features
- Added activity insights and share cards to the Profile section. You can review Codex usage highlights and save a profile card; sharing is available on consumer ChatGPT plans.
Performance improvements and bug fixes
- Improved Computer Use startup readiness and appshot error reporting.
- Fixed browser and review UI issues, including fullscreen browser composer controls, hex color swatches, terminal scrollbar alignment, and animated diff stat alignment.
- Expanded onboarding with more role choices so Codex can tailor first-run suggestions more accurately.
- Fixed configuration writes after plugin installation.
- Additional performance improvements and bug fixes.
-
Build and deploy websites with Sites
Sites is now available in preview in the Codex app. Use the Sites plugin to create, save, deploy, and inspect websites, dashboards, internal tools, web apps, and games hosted by OpenAI.
Open Sites in the app sidebar to return to your projects and manage hosted environment variables and secrets.
ChatGPT Business workspaces include Sites by default. ChatGPT Enterprise admins can enable Sites for the appropriate roles through role-based access control (RBAC).
-
ChatGPT for iOS 1.2026.146
New features
- Added an optional Face ID or passcode lock for Codex.
- Added a new settings screen for choosing Queue or Steer as the default follow-up behavior and toggling line wrapping for code diffs.
- Added support for connecting to Windows machines over SSH.
Improvements and bug fixes
- Added support for
/side <prompt>to start a side conversation with an initial question. - Improved follow-up prompts, the Codex home screen, and viewing changed files.
- Fixed issues with reconnecting, archiving threads, loading tasks, and connecting to hosts.
-
Use Codex with Amazon Bedrock
Codex can now use supported OpenAI models available through Amazon Bedrock. Configure Amazon Bedrock as your model provider to run Codex locally with AWS-managed authentication, account controls, and billing.
-
Terminal placement controls 26.601
New features
- Added Default terminal location in General settings. When the bottom panel is enabled, choose whether the terminal shortcut and environment actions open terminal tabs in the bottom panel or the right panel.
Performance improvements and bug fixes
- Additional performance improvements and bug fixes.
May 2026
-
Computer use and mobile access on Windows 26.527
New features
- Computer Use now works on Windows. Codex can operate Windows desktop apps by seeing, clicking, and typing in the foreground while it works.
- Remote control now supports Windows devices. You can start Codex work on a Windows device from ChatGPT on iOS or Android, or from a Mac running Codex, and check its progress remotely.
- The Profile section now shows your profile details, usage stats, and token activity.
- Added thread coordination for local projects and worktrees, including separate background threads when explicitly requested.
- Expanded search for past Codex app threads to include conversation content and Git branch names.
- Added stable identicons for background subagents across the app.
- Improved keyboard shortcut settings with keypress search and a reset-all action.
- Improved Chrome context capture for Google Docs, Sheets, and Slides tabs.
Performance improvements and bug fixes
- Additional performance improvements and bug fixes.
-
GPT-5.3-Codex and GPT-5.2 deprecated
GPT-5.3-Codex and GPT-5.2 are now deprecated as user-selectable models in Codex for users signed in with ChatGPT. API-key workflows aren’t affected.
Use a current Codex model, such as GPT-5.5, GPT-5.4, or GPT-5.4 mini. See Codex models for model availability and Codex pricing for credit rates.
-
ChatGPT for iOS 1.2026.139
New features
- Added Spotlight and Shortcuts support for opening Codex Mobile directly.
- Added browsing for archived Codex threads.
- Added
/sidefor opening a side conversation. - Added options to save or copy rendered images.
Improvements and bug fixes
- Improved iPad keyboard shortcuts.
- Improved setup and relaunch reliability.
- Fixed issues with task progress, loading archived threads, previewing code changes, and switching hosts.
-
Appshots, goal mode, and more 26.519
Appshots are now available in the Codex app on macOS. Press both Command keys to send the frontmost app window to Codex with a screenshot and available text, so Codex can work from context in another app without you copying, pasting, or describing it manually.
This launch also includes:
- Goal mode is no longer an experimental feature and is available in the Codex app, IDE extension, and CLI. With Goal mode, you can have Codex drive toward a specific objective for hours or even days.
- Remote computer use, so Codex can use desktop apps after your Mac locks, including remotely via Codex Mobile. Codex scopes locked use to active, trusted computer use turns and includes safeguards such as short-lived authorization, covered displays, relock on local input, and manual-unlock fallback.
- Plugin sharing through marketplace sources is available for ChatGPT Business. Enterprise support is coming soon. Teams can distribute reusable plugin bundles that include skills, MCP servers, and lifecycle hooks.
- Advanced in-app browser annotations let you tweak styling such as font size, colors, and spacing directly using annotations. This gives Codex a clearer signal for changes.
- Browser-use improvements across in-app browser & Chrome:
- Codex can now download and extract all image assets from a page much more quickly.
- Codex can now extract structured data from pages more effectively and find information more quickly with a read-only JS sandbox.
- Chrome extension will create less clutter when using it. Codex will no longer create tab groups when taking over existing tabs, and at the end of a task for handoff. Instead, it uses tab icons to indicate status.
- Significantly improved reliability for browser use. We fixed bugs on Windows, flaky availability of the plugin to non geo-blocked regions, and many other issues impacting performance.
-
ChatGPT for iOS 1.2026.132
New features
- Added support for opening completed Codex tasks directly from iOS notifications.
- Added the ability to open changed files directly while reviewing a task.
Improvements and bug fixes
- Improved task resume, reconnection, and foreground reliability.
- Improved task progress updates, code review, and message composition.
-
Work with Codex from anywhere
You can now use Codex from the ChatGPT mobile app by connecting it to a Mac running the Codex app. Codex runs from the connected host, so the same projects, files, credentials, plugins, skills, and configuration are available from your phone.
See Remote connections for mobile setup, choosing a host, what comes from the connected machine, and SSH hosts. This launch also includes Hooks general availability, Codex access tokens for trusted automation, and Enterprise admin setup guidance.
-
Expanded Auto-review documentation
Added a dedicated Auto-review page covering the reviewer lifecycle, trigger conditions, failure behavior, and local or managed configuration.
Also updated the Agent approvals & security and Sandbox docs so they explain more clearly how Auto-review relates to the sandbox boundary.
-
Codex app 26.506
New features
- Added an in-app trust review flow for hooks and kept Hooks settings reachable even before hooks are fully configured.
Performance improvements and bug fixes
- Restored tooltip-wrapped dropdowns that could stop opening after the tooltip rewrite.
- Preserved in-progress message edits across thread switches.
- Fixed several desktop workflow regressions, including
Ctrl+Vpaste in the Windows terminal, opening modified external links outside the in-app browser, and keeping feedback slash commands attached to the right thread. - Improved loading and panel polish by showing model loading while a thread resumes, hiding unavailable model controls during load, and bundling summary-panel layout and hover fixes.
- Kept the Computer Use settings control visible even when uninstalled and disabled problematic extension hover panels.
- Additional performance improvements and bug fixes.
-
Codex for Chrome
With the new extension for Chrome, Codex is even better at working with apps and websites in your browser. It works in parallel across tabs in the background without taking over your browser, and you stay in control of which websites Codex can use.
Learn more in the Codex Chrome extension documentation.
-
Codex analytics governance docs update
Updated the Codex enterprise governance guide with more detailed coverage of the Analytics dashboard charts, data export options, and enterprise Analytics API endpoints.
-
Create Codex access tokens
ChatGPT Enterprise workspace owners and admins can allow permitted members to create Codex access tokens for trusted, non-interactive Codex local workflows. Members can use access tokens to run Codex from scripts, schedulers, and private CI runners with their ChatGPT workspace identity.
Learn more in Access tokens.
-
Codex app 26.429
New features
- Added dictation cleanup plus a configurable dictation dictionary for names, file paths, and code symbols.
- Added zoom and download controls to the image lightbox.
Performance improvements and bug fixes
- Improved voice and dictation error messages for microphone, connection, and quota failures.
- Fixed in-app browser comment markers so they stay aligned across scrolling, zoom, and responsive layout changes.
- Made pull request creation and recovery flows more reliable by preserving newly created pull request state, classifying more app-server failures as restart-required, and stopping exhausted remote reconnect loops.
- Additional performance improvements and bug fixes.
April 2026
-
Codex app 26.423
New features
- Added a tooltip on realtime delegation messages to clarify that Codex uses the surrounding voice conversation as context.
Performance improvements and bug fixes
- Fixed search in long review files so next and previous results reliably jump to off-screen matches.
- Kept embedded MCP app panels from restarting or losing state during fullscreen changes and thread reloads.
- Fixed several desktop regressions, including tray crashes when the local connection is missing, duplicate macOS fullscreen menu entries, and broken global dictation hotkeys on older macOS versions.
- Additional performance improvements and bug fixes.
-
GPT-5.5 and Codex app updates
GPT-5.5 is now available in Codex as OpenAI’s newest frontier model for complex coding, computer use, knowledge work, and research workflows.
GPT-5.5 in Codex
GPT-5.5 is the recommended choice for most Codex tasks when it appears in your model picker. It’s especially useful for implementation, refactors, debugging, testing, validation, and knowledge-work artifacts.
To switch to GPT-5.5:
- In the CLI, start a new thread with:
Or usecodex --model gpt-5.5/modelduring a session. - In the IDE extension, choose GPT-5.5 from the model selector in the composer.
- In the Codex app, choose GPT-5.5 from the model selector in the composer.
If you don’t see GPT-5.5 yet, update the CLI, IDE extension, or Codex app to the latest version. During the rollout, continue using GPT-5.4 if GPT-5.5 is not yet available.
Browser use in the Codex app
The Codex app can now let Codex operate the in-app browser for local development servers and file-backed pages. Ask Codex to use the browser when it needs to click through a rendered UI, reproduce a visual bug, or verify a local fix inside the app.
Browser use runs through the bundled Browser plugin. In settings, you can manage the plugin and review allowed or blocked websites.
Automatic approval reviews
Codex can route eligible approval prompts through an automatic reviewer agent before the request runs. When configured, the Codex app shows an automatic review item with the review status and risk level, so you can see whether the reviewer approved, denied, stopped, or timed out before deciding.
- In the CLI, start a new thread with:
-
Codex app 26.417
New features
- Added local branch search and non-image file pasting in the composer.
- Added collapsible sidebar sections, tray usage-limit surfacing, and a command-palette theme switcher.
Performance improvements and bug fixes
- Made review faster and more stable with better diff batching and preserved diff and search state.
- Fixed projectless cwd and permissions handling, default file opening, spreadsheet suggestions, and remote-control reconnect issues.
- Additional performance improvements and bug fixes.
-
Codex can now help with more of your work 26.415
Codex is becoming a broader workspace for getting work done with AI. This update makes it easier to start work with less setup, verify what Codex is building, create richer outputs, and keep momentum across longer-running tasks.
Verify more of your work
The Codex app now includes an early in-app browser. You can open local or public pages that don’t require sign-in, comment directly on the rendered page, and ask Codex to address page-level feedback.
Computer use lets Codex operate macOS apps by seeing, clicking, and typing, which helps with native app testing, simulator flows, low-risk app settings, and GUI-only bugs.
The feature isn’t available in the European Economic Area, the United Kingdom, or Switzerland at launch.
Start, follow, and steer work
Chats are threads you can start without choosing a project folder first. They’re useful for research, writing, planning, analysis, source gathering, and tool-driven work that doesn’t begin in a codebase.
For work that needs a later check-in, thread automations can wake up the same thread on a schedule while preserving the conversation context. Use them to check a long-running process, watch for updates, or continue a follow-up loop without starting from scratch.
The task sidebar makes plans, sources, generated artifacts, and summaries easier to follow while Codex works. Context-aware suggestions can also help you pick up relevant follow-ups when you start or return to Codex.
Stronger for software development
Codex now brings more of the pull request workflow into the app. You can inspect GitHub pull requests in the sidebar, review comments in the diff, review changed files, then ask Codex to explain feedback, make changes, check them, and keep the review moving.
Review richer outputs
The artifact viewer can preview generated files such as PDF files, spreadsheets, documents, and presentations in the sidebar before you commit or share them. Memories, where available, can also carry useful context from past tasks into future threads, including stable preferences, project conventions, and recurring work patterns.
Other features
- Remote connections - We’re gradually rolling out SSH remote connections in alpha
- Support for multiple terminals
- macOS menu bar and Windows system tray support
- Multi-window support
- Intel Mac support
- New plugins
- Improved thread and tool rendering
-
Codex app 26.410
New features
- Added command-menu file search, including
Cmd+Prouting into workspace file search. - Added rich previews in the sidebar file viewer for images, PDFs, and Markdown.
- Added terminal tabs per thread, a selected-text Ask Codex overlay, and a Help menu feedback entry.
Performance improvements and bug fixes
- Improved review diff whitespace handling and search highlighting.
- Fixed in-app browser address bar and external-open issues, plus several file viewer and side-panel bugs.
- Additional performance improvements and bug fixes.
- Added command-menu file search, including
-
Codex app 26.409
New features
- Added Windows Store updater support.
- Expanded pull request workflows with an activity timeline, PR-page commenting, and push choices in the push modal.
- Added workspace file tabs in the thread side panel, drag-and-drop tab reordering, run action editing, and a logout confirmation dialog.
Performance improvements and bug fixes
- Improved pull request board performance and comment flyouts.
- Improved update and navigation resilience, and fixed projectless visibility, unread-state, and pinned-row edge cases.
- Additional performance improvements and bug fixes.
-
Codex app 26.406
New features
- Added collapsible inline review comments and inline or detached review modes.
- Added a Git summary and Sources section in the thread side panel.
- Added a New Quick Chat command and local video embeds in the app.
Performance improvements and bug fixes
- Preserved thread scroll position per conversation and unread state across windows.
- Improved review refresh reliability, and fixed dictation loss, right-panel reset, and GitHub reconnect messaging.
- Additional performance improvements and bug fixes.
-
Codex model availability update
We’re updating model availability for users who sign in with ChatGPT. Starting April 7, the model picker no longer shows
gpt-5.2-codex,gpt-5.1-codex-mini,gpt-5.1-codex-max,gpt-5.1-codex,gpt-5.1, orgpt-5. On April 14, we’ll remove those models from Codex for ChatGPT sign-in.Users can still choose from
gpt-5.4,gpt-5.4-mini,gpt-5.3-codex, andgpt-5.2. ChatGPT Pro users can also choosegpt-5.3-codex-spark.To use another API-supported model in Codex, sign in with an API key or configure a model provider.
-
Codex app 26.325, 26.331, 26.401
New features
- Added workspace settings to the app.
- Added “Don’t ask again” handling and polish for custom MCP approval panels.
- Added native Windows updater support, including MSIX support, plus a Windows system tray menu so Codex can stay resident after the last window closes.
- Added app and file
@mentions in the automation composer, surfaced subagent diff stats in the composer, and added artifact cards for generated file citations. - Added a Quick Chat app-menu shortcut, a review file tree open menu, early heartbeat automation affordances in threads, and image support for remote connections.
Performance improvements and bug fixes
- Fixed review panel scroll jumps and PR status actions while a conversation is still running.
- Fixed several multi-window issues, plus
@-mention results, duplicate project labeling, WindowsrunGitbehavior, and revert, unstage, and stage-all actions. - Improved remote-thread and sidebar polish, Windows update recovery, unsupported-version guidance, and overall thread search speed.
- Fixed sticky review issues such as diff hunk expansion, header overlap, archive-thread crashes, and window-zoom shell sizing.
- Additional performance improvements and bug fixes.
March 2026
-
Build and install plugins in Codex
Codex now supports plugins: installable bundles that package skills, app integrations, and MCP server configuration for reusable workflows.
Plugins are available in the Codex app, CLI, and IDE extensions.
You can install curated plugins from the plugin directory, or scaffold a local plugin with
@plugin-creatorand test it with workspace-scoped or home-scoped marketplaces.Learn more in the plugins documentation.

Plugin structure
Every plugin is a folder with a required
.codex-plugin/plugin.jsonmanifest and optional supporting files:my-plugin/ .codex-plugin/ plugin.json # Required: plugin manifest skills/ # Optional: packaged skills .app.json # Optional: app or connector mappings .mcp.json # Optional: MCP server configuration assets/ # Optional: icons, logos, screenshotsInstall plugins per-user or per-repo
You can install plugins for just yourself with
~/.agents/plugins/marketplace.jsonand~/.codex/plugins/, or for everyone on a project with.agents/plugins/marketplace.jsonand a repo-local plugin directory such as./plugins/.Curated plugins and local development
Codex surfaces curated public plugins in the plugin directory. Codex also ships with the built-in
@plugin-creatorskill to help you scaffold a plugin, add a local marketplace entry, and test it before sharing it with teammates. -
Codex app 26.324
New features
- Redesigned the skills and plugins browse and manage pages.
- Added per-window zoom and a clearer edited-files state in review.
- Added automation titles and icons in the sidebar, plus bundled Raycast themes.
Performance improvements and bug fixes
- Kept loaded threads and projects visible during reconnects and made navigation feel faster.
- Fixed archive freezes, markdown wrapping, hotkey-window regressions, and several permissions, terminal, and worktree issues.
- Additional performance improvements and bug fixes.
-
Codex app 26.323
New features
- Added search for past Codex app threads, including a sidebar shortcut and keyboard shortcuts for jumping to recent threads.
- Added a one-click option to archive all local threads in a project.
- Synced key settings between the Codex app and the VS Code extension, and added a settings entry point in the extension.
Performance improvements and bug fixes
- Additional performance improvements and bug fixes.
-
Codex app 26.320
New features
- Added Floating Composer v2.
- Added terminal shortcuts for jumping by word and line.
- Improved plugin discovery surfaces and file-path rendering for saved images.
Performance improvements and bug fixes
- Fixed sidebar crashes when subagent turn items are missing.
- Fixed pop-out thread routing and preserved local paths for composer image attachments.
- Additional performance improvements and bug fixes.
-
Codex app 26.318, 26.319
New features
- Added skills to the
@menu so you can insert them from the composer alongside other mentions. Cmd/Ctrl+Fnow starts with your current text selection, which makes searching reviews and diffs faster, alongside broader review navigation improvements such as a refreshed file tree and percentage-based file tree resizing.- Added a branded loading shimmer while the app starts.
Performance improvements and bug fixes
- Improved collapsed diff summaries in review.
- Fixed slash-command focus and composer alignment issues, and polished plugin cards and step details.
- Additional performance improvements and bug fixes.
- Added skills to the
-
Codex app 26.317
New features
- You can now fork a conversation from an earlier message, not just the latest turn.
- Added slash commands for switching models and reasoning levels, and made slash commands work in the middle of a draft prompt.
- Added notifications for plan mode questions so it’s easier to notice when Codex needs input.
Performance improvements and bug fixes
- Fixed thread handoff and subagent navigation issues across worktrees and the VS Code extension.
- Additional performance improvements and bug fixes.
-
Introducing GPT-5.4 mini in Codex
GPT-5.4 mini is now available in Codex as a fast, efficient model for lighter coding tasks and subagents.
It improves over GPT-5 mini across coding, reasoning, image understanding, and tool use while running more than 2x faster. In Codex, GPT-5.4 mini uses 30% as much of your included limits as GPT-5.4, so comparable tasks can last about 3.3x longer before you hit those limits.
GPT-5.4 mini is available in the Codex app, the CLI, the IDE extension, and Codex on the web. GPT-5.4 mini is also available in the API.
Use GPT-5.4 mini for codebase exploration, large-file review, processing supporting documents, and other less reasoning-intensive subagent work. For more complex planning, coordination, and final judgment, start with GPT-5.4.
To switch to GPT-5.4 mini:
- In the CLI, start a new thread with:
Or usecodex --model gpt-5.4-mini/modelduring a session. - In the IDE extension, choose GPT-5.4 mini from the model selector in the composer.
- In the Codex app, choose GPT-5.4 mini from the model selector in the composer.
If you don’t see GPT-5.4 mini yet, update the CLI, IDE extension, or Codex app to the latest version.
- In the CLI, start a new thread with:
-
Codex app 26.313
New features
- Added back and forward buttons in the header so you can move between recent screens more quickly.
- Added an Open in Finder, Open in Explorer, or Open in File Manager action from thread menus to jump straight to a thread’s project folder.
Performance improvements and bug fixes
- Improved resume and thread error toasts with clearer details when something goes wrong.
- Additional performance improvements and bug fixes.
-
Codex app 26.312
Themes
Change the Codex app appearance in Settings by choosing a base theme, adjusting accent, background, and foreground colors, and changing the UI and code fonts. You can also share your custom theme with friends.
Revamped Automations
You can now choose whether automations run locally or on a worktree, define custom reasoning levels and models, and use templates to find inspiration for new automations.
Performance improvements and bug fixes
Various bug fixes and performance improvements.
-
Codex app 26.311
New features
- Codex can now read the integrated terminal for the current thread, so it can check the status of a running development server or refer back to failed build output while it works with you.
Performance improvements and bug fixes
- Additional performance improvements and bug fixes.
-
Introducing GPT-5.4 in Codex
GPT-5.4 is now available in Codex as OpenAI’s most capable and efficient frontier model for professional work.
It combines recent advances in reasoning, coding, and agentic workflows in one model, and it’s the recommended choice for most Codex tasks.
In Codex, GPT-5.4 is the first general-purpose model with native computer-use capabilities. GPT-5.4 in Codex includes experimental support for the 1M context window. It supports complex workflows across applications and long-horizon tasks, with stronger tool use and tool search that help agents find and use the right tools more efficiently.
GPT-5.4 is available everywhere you can use Codex: the Codex app, the CLI, the IDE extension, and Codex Cloud on the web. GPT-5.4 is also available in the API.
To switch to GPT-5.4:
- In the CLI, start a new thread with:
Or usecodex --model gpt-5.4/modelduring a session. - In the IDE extension, choose GPT-5.4 from the model selector in the composer.
- In the Codex app, choose GPT-5.4 from the model selector in the composer.
If you don’t see GPT-5.4 yet, update the CLI, IDE extension, or Codex app to the latest version.
- In the CLI, start a new thread with:
-
Codex app 26.305
Performance improvements and bug fixes
- Improved remote connections with clearer connection errors, better status updates, and clearer host labels in thread and settings views.
- Fixed copy and paste shortcuts in the integrated terminal on Windows.
- Fixed an issue where archived pinned threads could reappear in the sidebar.
- Fixed an issue where repeated
codex://newlinks could stop prefilling a new conversation when the app was already open. - Additional performance improvements and bug fixes.
-
Codex app 26.304
Codex app for Windows
The Codex app is now available on Windows. The app gives you one interface for working across projects, running parallel agent threads, and reviewing results in one place.
The Codex app runs natively on Windows using PowerShell and a native Windows sandbox for bounded permissions, so you can use Codex on Windows without moving your workflow into WSL, onto a virtual machine, or by deactivating the sandbox.
The Windows app includes the same core features as the rest of the Codex app:
- Skills to discover and extend Codex capabilities.
- Automations to run work in the background.
- Worktrees to handle independent tasks in the same project.
If you prefer to develop in WSL, you can also switch the Codex agent and the integrated terminal to run there.
Download it from the Microsoft Store and sign in with your ChatGPT account or an API key. For setup and configuration details, see Setup, Use WSL with the Codex app, and Customize the app for your development setup.
-
Codex app 26.303
New features
- Added a Worktrees setting to turn automatic cleanup of Codex-managed worktrees on or off.
- Added Handoff support for moving a thread between Local and Worktree.
- Added an explicit English option in the language menu.
Performance improvements and bug fixes
- Improved GitHub and pull request workflows.
- Improved approval prompts and app connection sign-in flows.
- Additional performance improvements and bug fixes.
February 2026
-
Codex app 26.228
Performance improvements and bug fixes
- Fixed a regression where conversation and task views could stop updating while Codex was streaming a response.
- Additional performance improvements and bug fixes.
-
Codex app 26.227
New features
- Added pull request status badges in task rows and PR buttons, including draft, open, merged, and closed states.
- Added a Worktrees setting to choose how many Codex-managed worktrees to keep before older ones are cleaned up.
Performance improvements and bug fixes
- Improved scrolling and navigation in long conversations and code review, including fixes for thread jumpiness, sidebar jitter, and diff scrolling.
- Improved app startup reliability and keyboard zoom behavior.
- Additional performance improvements and bug fixes.
-
Codex app 26.226
New features
- Added new MCP shortcuts in the composer, including install keyword suggestions and an MCP server submenu in Add context.
- Added support for
@mentionsand skill mentions in inline review comments.
Performance improvements and bug fixes
- Improved rendering of MCP tool calls and Mermaid diagram error handling.
- Fixed an issue where stopped terminal commands could continue appearing as running.
- Additional performance improvements and bug fixes.
-
Codex app 26.217
New features
- Added drag-and-drop support to reorder queued messages.
- Added a warning when the selected model is downgraded.
Improvements and bug fixes
- Improved file workflows with fuzzy file search and better attachment recovery after restart.
- Additional performance improvements and bug fixes.
-
Introducing GPT-5.3-Codex-Spark
Today, we’re releasing a research preview of GPT-5.3-Codex-Spark, a smaller version of GPT-5.3-Codex and our first model designed for real-time coding. Codex-Spark is optimized to feel near-instant, delivering more than 1000 tokens per second while remaining highly capable for real-world coding tasks.
Codex-Spark is available in research preview for ChatGPT Pro users in the latest Codex app, CLI, and IDE extension. This release also marks the first milestone in our partnership with Cerebras.
At launch, Codex-Spark is text-only with a 128k context window. During the research preview, usage has separate model-specific limits and doesn’t count against standard Codex limits. During high demand, access may slow down or queue while we balance reliability across users.
To switch to GPT-5.3-Codex-Spark:
- In the CLI, start a new thread with:
Or usecodex --model gpt-5.3-codex-spark/modelduring a session. - In the IDE extension, choose GPT-5.3-Codex-Spark from the model selector in the composer.
- In the Codex app, choose GPT-5.3-Codex-Spark from the model selector in the composer.
If you don’t see GPT-5.3-Codex-Spark yet, update the CLI, IDE extension, or Codex app to the latest version.
GPT-5.3-Codex-Spark isn’t available in the API at launch. For API-key workflows, continue using
gpt-5.2-codex. - In the CLI, start a new thread with:
-
Codex app 26.212
New features
- Support for GPT-5.3-Codex-Spark
- Added conversation forking
- Added floating pop-out window to take a conversation with you
Bug fixes
- Improved performance and bug fixes
Alpha testing for the Codex app on Windows is also starting. Sign up here to be a potential alpha tester.
-
Codex app 26.210
New features
- Added branch search in the branch picker.
- Added clearer guidance for entering plan mode when you type
planin the composer. - Added support for parallel approvals.
Improvements and bug fixes
- Additional performance improvements and bug fixes.
-
GPT-5.3-Codex in Cursor and VS Code
Starting today, GPT-5.3-Codex is available natively in Cursor and VS Code.
API access is starting with a small set of customers as part of a phased release.
This is the first model treated as a high security capability under the Preparedness Framework.
Safety controls will continue to scale, and API access will expand over the next few weeks.
-
Codex app 26.208
New features
- Added MCP and personality actions to the command palette.
- Updated follow-up behavior to queue by default.
Improvements and bug fixes
- Additional performance improvements and bug fixes.
-
Codex app 26.206
New features
- Added a file-reference action to reveal files directly in your OS file manager.
Improvements and bug fixes
- Improved handling of large reviews by removing the overall diff-size cap in the review pane.
- Additional performance improvements and bug fixes.
-
Introducing GPT-5.3-Codex
Today we’re releasing GPT-5.3-Codex, the most capable agentic coding model to date for complex, real-world software engineering.
GPT-5.3-Codex combines the frontier coding performance of GPT-5.2-Codex with stronger reasoning and professional knowledge capabilities, and runs 25% faster for Codex users. It’s also better at collaboration while the agent is working—delivering more frequent progress updates and responding to steering in real time.
GPT-5.3-Codex is available with paid ChatGPT plans everywhere you can use Codex: the Codex app, the CLI, the IDE extension, and Codex Cloud on the web. API access for the model will come soon.
To switch to GPT-5.3-Codex:
- In the CLI, start a new thread with:
Or usecodex --model gpt-5.3-codex/modelduring a session. - In the IDE extension, make sure you are signed in with ChatGPT, then choose GPT-5.3-Codex from the model selector in the composer.
- In the Codex app, make sure you are signed in with ChatGPT, then choose GPT-5.3-Codex from the model selector in the composer.
- If you don’t see GPT-5.3-Codex, update the CLI, IDE extension, or Codex app to the latest version.
For API-key workflows, continue using
gpt-5.2-codexwhile API support rolls out. - In the CLI, start a new thread with:
-
Codex app 26.205
New features
- Support for GPT-5.3-Codex.
- Added mid-turn steering. Submit a message while Codex is working to direct its behavior.
- Attach or drop any file type.
Bug fixes
- Fix flickering of the app.
-
Codex app 26.204
New features
- Added Zed and Textmate as options to open files and folders.
- Added PDF preview in the review panel.
Bug fixes
- Performance improvements.
-
Codex app 26.203
New features
- Added thread renaming on double-click in the thread list.
Improvements and bug fixes
- Renamed Sync to Handoff and added clearer source/destination stats in the handoff UI.
- Additional performance improvements and bug fixes.
-
Introducing the Codex app
Codex app
The Codex app for macOS is a desktop interface for running agent threads in parallel and collaborating with agents on long-running tasks. It includes a project sidebar, thread list, and review pane for tracking work across projects.
Key features:
- Multitask across projects
- Built-in worktree support
- Voice dictation
- Built-in Git tooling
- Skills
- Automations
For a limited time, ChatGPT Free and Go include Codex, and Plus, Pro, Business, Enterprise, and Edu plans get double rate limits. Those higher limits apply in the app, the CLI, your IDE, and the cloud.
Learn more in the Introducing the Codex app blog post.
Check out the Codex app documentation for more.
January 2026
-
Web search is now enabled by default
Codex now enables web search for local tasks in the Codex CLI and IDE Extension. By default, Codex uses a web search cache, which is an OpenAI-maintained index of web results. Cached mode returns pre-indexed results instead of fetching live pages, while live mode fetches the most recent data from the web. If you are using
--yoloor another full access sandbox setting, web search defaults to live results. To disable this behavior or switch modes, use theweb_searchconfiguration option:web_search = "cached"(default; serves results from the web search cache)web_search = "live"(fetches the most recent data from the web; same as--search)web_search = "disabled"to remove the tool
To learn more, check out the configuration documentation.
-
Team Config for shared configuration
Team Config groups the files teams use to standardize Codex across repositories and machines. Use it to share:
config.tomldefaultsrules/for command controls outside the sandboxskills/for reusable workflows
Codex loads these layers from
.codex/folders in the current working directory, parent folders, and the repo root, plus user (~/.codex/) and system (/etc/codex/) locations. Higher-precedence locations override lower-precedence ones.Admins can still enforce constraints with
requirements.toml, which overrides defaults regardless of location.Learn more in Team Config.
-
Custom prompts deprecated
Custom prompts are now deprecated. Use skills for reusable instructions and workflows instead.
-
GPT-5.2-Codex API availability
GPT-5.2-Codex is now available in the API and for users who sign into Codex with the API.
To learn more about using GPT-5.2-Codex check out our API documentation.
December 2025
-
Agent skills in Codex
Codex now supports agent skills: reusable bundles of instructions (plus optional scripts and resources) that help Codex reliably complete specific tasks.
Skills are available in both the Codex CLI and IDE extensions.
You can invoke a skill explicitly by typing
$skill-name(for example,$skill-installeror the experimental$create-planskill after installing it), or let Codex select a skill automatically based on your prompt.Learn more in the skills documentation.




Folder-based standard (agentskills.io)
Following the open agent skills specification, a skill is a folder with a required
SKILL.mdand optional supporting files:my-skill/ SKILL.md # Required: instructions + metadata scripts/ # Optional: executable code references/ # Optional: documentation assets/ # Optional: templates, resourcesInstall skills per-user or per-repo
You can install skills for just yourself in
~/.codex/skills, or for everyone on a project by checking them into.codex/skillsin the repository.Codex also ships with a few built-in system skills to get started, including
$skill-creatorand$skill-installer. The$create-planskill is experimental and needs to be installed (for example:$skill-installer install the create-plan skill from the .experimental folder).Curated skills directory
Codex ships with a small curated set of skills inspired by popular workflows at OpenAI. Install them with
$skill-installer, and expect more over time. -
Introducing GPT-5.2-Codex
Today we are releasing GPT-5.2-Codex, the most advanced agentic coding model yet for complex, real-world software engineering.
GPT-5.2-Codex is a version of GPT-5.2 further optimized for agentic coding in Codex, including improvements on long-horizon work through context compaction, stronger performance on large code changes like refactors and migrations, improved performance in Windows environments, and significantly stronger cybersecurity capabilities.
Starting today, the CLI and IDE Extension will default to
gpt-5.2-codexfor users who are signed in with ChatGPT. API access for the model will come soon.If you have a model specified in your
config.tomlconfiguration file, you can instead try outgpt-5.2-codexfor a new Codex CLI session using:codex --model gpt-5.2-codexYou can also use the
/modelslash command in the CLI. In the Codex IDE Extension you can select GPT-5.2-Codex from the dropdown menu.If you want to switch for all sessions, you can change your default model to
gpt-5.2-codexby updating yourconfig.tomlconfiguration file:model = "gpt-5.2-codex” -
Introducing Codex for Linear
Assign or mention @Codex in an issue to kick-off a Codex cloud task. As Codex works, it posts updates back to Linear, providing a link to the completed task so you can review, open a PR, or keep working.

To learn more about how to connect Codex to Linear both locally through MCP and through the new integration, check out the Codex for Linear documentation.
November 2025
-
Usage and credits fixes
Minor updates to address a few issues with Codex usage and credits:
- Adjusted all usage dashboards to show “limits remaining” for consistency. The CLI previously displayed “limits used.”
- Fixed an issue preventing users from buying credits if their ChatGPT subscription was purchased via iOS or Google Play.
- Fixed an issue where the CLI could display stale usage information; it now refreshes without needing to send a message first.
- Optimized the backend to help smooth out usage throughout the day, irrespective of overall Codex load or how traffic is routed. Before, users could get unlucky and hit a few cache misses in a row, leading to much less usage.
-
Introducing GPT-5.1-Codex-Max
Today we are releasing GPT-5.1-Codex-Max, our new frontier agentic coding model.
GPT‑5.1-Codex-Max is built on an update to our foundational reasoning model, which is trained on agentic tasks across software engineering, math, research, and more. GPT‑5.1-Codex-Max is faster, more intelligent, and more token-efficient at every stage of the development cycle–and a new step towards becoming a reliable coding partner.
Starting today, the CLI and IDE Extension will default to
gpt-5.1-codex-maxfor users that are signed in with ChatGPT. API access for the model will come soon.For non-latency-sensitive tasks, we’ve also added a new Extra High (
xhigh) reasoning effort, which lets the model think for an even longer period of time for a better answer. We still recommend medium as your daily driver for most tasks.If you have a model specified in your
config.tomlconfiguration file, you can instead try outgpt-5.1-codex-maxfor a new Codex CLI session using:codex --model gpt-5.1-codex-maxYou can also use the
/modelslash command in the CLI. In the Codex IDE Extension you can select GPT-5.1-Codex from the dropdown menu.If you want to switch for all sessions, you can change your default model to
gpt-5.1-codex-maxby updating yourconfig.tomlconfiguration file:model = "gpt-5.1-codex-max” -
Introducing GPT-5.1-Codex and GPT-5.1-Codex-Mini
Along with the GPT-5.1 launch in the API, we are introducing new
gpt-5.1-codex-miniandgpt-5.1-codexmodel options in Codex, a version of GPT-5.1 optimized for long-running, agentic coding tasks and use in coding agent harnesses in Codex or Codex-like harnesses.Starting today, the CLI and IDE Extension will default to
gpt-5.1-codexon macOS and Linux andgpt-5.1on Windows.If you have a model specified in your
config.tomlconfiguration file, you can instead try outgpt-5.1-codexfor a new Codex CLI session using:codex --model gpt-5.1-codexYou can also use the
/modelslash command in the CLI. In the Codex IDE Extension you can select GPT-5.1-Codex from the dropdown menu.If you want to switch for all sessions, you can change your default model to
gpt-5.1-codexby updating yourconfig.tomlconfiguration file:model = "gpt-5.1-codex” -
Introducing GPT-5-Codex-Mini
Today we are introducing a new
gpt-5-codex-minimodel option to Codex CLI and the IDE Extension. The model is a smaller, more cost-effective, but less capable version ofgpt-5-codexthat provides approximately 4x more usage as part of your ChatGPT subscription.Starting today, the CLI and IDE Extension will automatically suggest switching to
gpt-5-codex-miniwhen you reach 90% of your 5-hour usage limit, to help you work longer without interruptions.You can try the model for a new Codex CLI session using:
codex --model gpt-5-codex-miniYou can also use the
/modelslash command in the CLI. In the Codex IDE Extension you can select GPT-5-Codex-Mini from the dropdown menu.Alternatively, you can change your default model to
gpt-5-codex-miniby updating yourconfig.tomlconfiguration file:model = "gpt-5-codex-mini” -
GPT-5-Codex model update
We’ve shipped a minor update to GPT-5-Codex:
- More reliable file edits with
apply_patch. - Fewer destructive actions such as
git reset. - More collaborative behavior when encountering user edits in files.
- 3% more efficient in time and usage.
- More reliable file edits with
October 2025
-
Credits on ChatGPT Pro and Plus
Codex users on ChatGPT Plus and Pro can now use on-demand credits for more Codex usage beyond what’s included in your plan. Learn more.
-
Tag @Codex on GitHub Issues and PRs
You can now tag
@codexon a teammate’s pull request to ask clarifying questions, request a follow-up, or ask Codex to make changes. GitHub Issues now also support@codexmentions, so you can kick off tasks from any issue, without leaving your workflow.
-
Codex is now GA
Codex is now generally available with 3 new features — @Codex in Slack, Codex SDK, and new admin tools.
@Codex in Slack

You can now questions and assign tasks to Codex directly from Slack. See the Slack guide to get started.
Codex SDK
Integrate the same agent that powers the Codex CLI inside your own tools and workflows with the Codex SDK in Typescript. With the new Codex GitHub Action, you can easily add Codex to CI/CD workflows. See the Codex SDK guide to get started.
import { Codex } from "@openai/codex-sdk"; const agent = new Codex(); const thread = await agent.startThread(); const result = await thread.run("Explore this repo"); console.log(result); const result2 = await thread.run("Propose changes"); console.log(result2);New admin controls and analytics

ChatGPT workspace admins can now edit or delete Codex Cloud environments. With managed config files, they can set safe defaults for CLI and IDE usage and monitor how Codex uses commands locally. New analytics dashboards help you track Codex usage and code review feedback. Learn more in the enterprise admin guide.
Availability and pricing updates
The Slack integration and Codex SDK are available to developers on ChatGPT Plus, Pro, Business, Edu, and Enterprise plans starting today, while the new admin features will be available to Business, Edu, and Enterprise. Beginning October 20, Codex Cloud tasks will count toward your Codex usage. Review the Codex pricing guide for plan-specific details.
September 2025
-
GPT-5-Codex in the API
GPT-5-Codex is now available in the Responses API, and you can also use it with your API Key in the Codex CLI. We plan on regularly updating this model snapshot. It is available at the same price as GPT-5. You can learn more about pricing and rate limits for this model on our model page.
-
Introducing GPT-5-Codex
New model: GPT-5-Codex

GPT-5-Codex is a version of GPT-5 further optimized for agentic coding in Codex. It’s available in the IDE extension and CLI when you sign in with your ChatGPT account. It also powers the cloud agent and Code Review in GitHub.
To learn more about GPT-5-Codex and how it performs compared to GPT-5 on software engineering tasks, see our announcement blog post.
Image outputs

When working in the cloud on front-end engineering tasks, GPT-5-Codex can now display screenshots of the UI in Codex web for you to review. With image output, you can iterate on the design without needing to check out the branch locally.
New in Codex CLI
- You can now resume sessions where you left off with
codex resume. - Context compaction automatically summarizes the session as it approaches the context window limit.
Learn more in the latest release notes
- You can now resume sessions where you left off with
August 2025
-
Late August update
IDE extension (Compatible with VS Code, Cursor, Windsurf)
Codex now runs in your IDE with an interactive UI for fast local iteration. Easily switch between modes and reasoning efforts.
Sign in with ChatGPT (IDE & CLI)
One-click authentication that removes API keys and uses ChatGPT Enterprise credits.
Move work between local ↔ cloud
Hand off tasks to Codex web from the IDE with the ability to apply changes locally so you can delegate jobs without leaving your editor.
Code Reviews
Codex goes beyond static analysis. It checks a PR against its intent, reasons across the codebase and dependencies, and can run code to validate the behavior of changes.
-
Mid August update
Image inputs

You can now attach images to your prompts in Codex web. This is great for asking Codex to implement frontend changes or follow up on whiteboarding sessions.
Container caching

Codex now caches containers to start new tasks and followups 90% faster, dropping the median start time from 48 seconds to 5 seconds. You can optionally configure a maintenance script to update the environment from its cached state to prepare for new tasks. See the docs for more.
Automatic environment setup
Now, environments without manual setup scripts automatically run the standard installation commands for common package managers like yarn, pnpm, npm, go mod, gradle, pip, poetry, uv, and cargo. This reduces test failures for new environments by 40%.
June 2025
-
Best of N

Codex can now generate multiple responses simultaneously for a single task, helping you quickly explore possible solutions to pick the best approach.
Fixes & improvements
Added some keyboard shortcuts and a page to explore them. Open it by pressing ⌘-/ on macOS and Ctrl+/ on other platforms.
Added a “branch” query parameter in addition to the existing “environment”, “prompt” and “tab=archived” parameters.
Added a loading indicator when downloading a repo during container setup.
Added support for cancelling tasks.
Fixed issues causing tasks to fail during setup.
Fixed issues running followups in environments where the setup script changes files that are gitignored.
Improved how the agent understands and reacts to network access restrictions.
Increased the update rate of text describing what Codex is doing.
Increased the limit for setup script duration to 20 minutes for Pro and Business users.
Polished code diffs: You can now option-click a code diff header to expand/collapse all of them.
-
June update
Agent internet access

Now you can give Codex access to the internet during task execution to install dependencies, upgrade packages, run tests that need external resources, and more.
Internet access is off by default. Plus, Pro, and Business users can enable it for specific environments, with granular control of which domains and HTTP methods Codex can access. Internet access for Enterprise users is coming soon.
Learn more about usage and risks in the docs.
Update existing PRs

Now you can update existing pull requests when following up on a task.
Voice dictation
Now you can dictate tasks to Codex.
Fixes & improvements
Added a link to this changelog from the profile menu.
Added support for binary files: When applying patches, all file operations are supported. When using PRs, only deleting or renaming binary files is supported for now.
Fixed an issue on iOS where follow up tasks where shown duplicated in the task list.
Fixed an issue on iOS where pull request statuses were out of date.
Fixed an issue with follow ups where the environments were incorrectly started with the state from the first turn, rather than the most recent state.
Fixed internationalization of task events and logs.
Improved error messages for setup scripts.
Increased the limit on task diffs from 1 MB to 5 MB.
Increased the limit for setup script duration from 5 to 10 minutes.
Polished GitHub connection flow.
Re-enabled Live Activities on iOS after resolving an issue with missed notifications.
Removed the mandatory two-factor authentication requirement for users using SSO or social logins.
May 2025
-
Reworked environment page
It’s now easier and faster to set up code execution.

Fixes & improvements
Added a button to retry failed tasks
Added indicators to show that the agent runs without network access after setup
Added options to copy git patches after pushing a PR
Added support for unicode branch names
Fixed a bug where secrets were not piped to the setup script
Fixed creating branches when there’s a branch name conflict.
Fixed rendering diffs with multi-character emojis.
Improved error messages when starting tasks, running setup scripts, pushing PRs, or disconnected from GitHub to be more specific and indicate how to resolve the error.
Improved onboarding for teams.
Polished how new tasks look while loading.
Polished the followup composer.
Reduced GitHub disconnects by 90%.
Reduced PR creation latency by 35%.
Reduced tool call latency by 50%.
Reduced task completion latency by 20%.
Started setting page titles to task names so Codex tabs are easier to tell apart.
Tweaked the system prompt so that agent knows it’s working without network, and can suggest that the user set up dependencies.
Updated the docs.
-
Codex in the ChatGPT iOS app
Start tasks, view diffs, and push PRs—while you’re away from your desk.
