Choosing a model is about finding the right balance of speed, cost, and capability for your task. Sometimes you need a quick first draft. Other times, a complex problem calls for more careful reasoning. This guide introduces our latest models, shows where each one shines, and helps you find a starting point for your work.
GPT-5.5 retirement: See replacement models and migration steps.
Choose a model
In Work or Codex in the ChatGPT desktop app, use the model and reasoning control beneath the composer to choose an available model and adjust its reasoning effort.
Reasoning effort guides how much work the model puts into thinking through your request. Higher effort can improve results for complex tasks, but it takes longer and uses more tokens. Start with the default effort and increase it when the task needs deeper planning or analysis.
Ultra mode goes beyond a single-agent run. It uses subagents to accelerate complex work, making it useful for larger tasks that can be split across subagents.
Choose a model
These recommendations apply to ChatGPT Work on the web. Use the model and reasoning control beneath the composer to choose an available model and adjust its reasoning effort.
Reasoning effort guides how much work the model puts into thinking through your request. Higher effort can improve results for complex tasks, but it takes longer and uses more tokens. Start with the default effort and increase it when the task needs deeper planning or analysis.
Ultra mode goes beyond a single-agent run. It uses subagents to accelerate complex work, making it useful for larger tasks that can be split across subagents.
Choose a model
GPT-6.1 Terra requires Codex CLI 0.159.0 or later. To use it, run
codex -m gpt-6.1-terra.
In an interactive CLI session, use /model to switch models or adjust
reasoning effort. You can also choose a model when you launch Codex with
--model or its -m alias:
codex --model gpt-6.1-solThe same option works with non-interactive runs. For example:
codex exec -m gpt-6.1-sol "Review the current changes"Reasoning effort guides how much work the model puts into thinking through your request. Higher effort can improve results for complex tasks, but it takes longer and uses more tokens. Start with the default effort and increase it when the task needs deeper planning or analysis.
Ultra mode goes beyond a single-agent run. It uses subagents to accelerate complex work, making it useful for larger tasks that can be split across subagents.
Choose a model
Use the model switcher below the composer to choose an available model and reasoning effort.
Reasoning effort guides how much work the model puts into thinking through your request. Higher effort can improve results for complex tasks, but it takes longer and uses more tokens. Start with the default effort and increase it when the task needs deeper planning or analysis.
Ultra mode goes beyond a single-agent run. It uses subagents to accelerate complex work, making it useful for larger tasks that can be split across subagents.
In the ChatGPT desktop app and ChatGPT Work on the web, start with the default Power setting available to your account. Move toward Smarter for deeper reasoning or Faster for faster, lower-cost work. Open Advanced to choose a specific model, reasoning effort, or speed. To try GPT-6.1 Terra in this example, open Advanced and choose it from the model list.
The six Power presets shown are Luna High, GPT-6 Sol Light (selected in this example), GPT-6 Sol Medium, Astra Light, Astra Medium, and Astra Extra High. Some paid plans omit Astra Extra High. Advanced controls are illustrative; available options and defaults vary by plan, client, workspace settings, and rollout.
Recommended models
Start with the kind of work you want to do. Astra suits the most demanding workflows. GPT-6.1 Sol offers near-Astra performance at a lower cost than Astra. GPT-6.1 Terra is a fast, efficient choice for writing, visual understanding, and computer use. Luna suits focused, repeatable tasks.
These recommendations apply to ChatGPT Work and Codex. For Chat, see models in ChatGPT in the Help Center.
Our most capable model for complex work across code, apps, and research, combining advanced reasoning, computer use, and stronger judgment.
Near-Astra performance for complex work at a lower cost than Astra. Consider it for repeated, long-running work across code, apps, and documents.
A fast, efficient model for clear writing, visual understanding, and computer use. Try it for drafting proposals, interpreting charts, or working with information in an app.
Our most efficient model for focused, high-volume tasks, including summarization, extraction, and focused coding.
For model availability, pricing, and usage limits, see ChatGPT Work and Codex in the Help Center.
Choosing a model
Start with a task you know well and compare the result with what you need. Consider how long you’re willing to wait, how much usage the task requires, and the level of detail or judgment involved. You can change models as your work changes.
Where each model shines
- Astra, for the hardest end-to-end work. Choose Astra for complete workflows across code, apps, and research that need sustained reasoning and judgment. Give it the sources, templates, constraints, and checks that define a useful result. Astra is better at asking focused questions and incorporating your guidance while keeping the original goal and constraints in view.
- GPT-6.1 Sol, for repeated, long-running work. Consider GPT-6.1 Sol for work across code, apps, and documents when cost matters. Keep Astra for your most demanding work.
- GPT-6.1 Terra, for everyday writing and visual work. Choose Terra for clear communication, interpreting images and screens, and completing tasks across supported apps. Try it for drafting proposals, reviewing visual materials, or filling out forms from source documents.
- Luna, for clear, repeatable tasks. Choose Luna for specific, high-volume tasks when you know what a good result looks like, such as extraction, classification, transformation, and structured summaries.
Try an example
Use your own materials with these prompts, and tell ChatGPT what a useful result should include.
| Model | Task | Example prompt |
|---|---|---|
| Astra | Work through a complex decision. | “Compare these three vendor proposals against our requirements. Explain the tradeoffs, recommend an option, and list the questions we should resolve before deciding.” |
| GPT-6.1 Sol | Bring several sources into a working document. | “Use this project brief, meeting notes, and timeline to draft an implementation plan. Identify dependencies, flag conflicting details, and list the decisions we still need to make.” |
| GPT-6.1 Terra | Make sense of visual information. | “Review these dashboard screenshots and summarize the main changes. Refer to the chart labels and values, and flag anything you can’t read confidently.” |
| Luna | Organize information into a consistent format. | “Turn these meeting notes into a table of action items, owners, and due dates. Mark any missing owner or date as unspecified.” |
Review the result against your source materials, especially names, numbers, and details taken from images. For tasks on websites or in desktop apps, check the browser use and computer use setup requirements first.
Pick a reasoning effort
For example, you might use a lower effort to rewrite a short email and a higher effort to compare several proposals with competing requirements.
For GPT-6.1 Sol, start with the reasoning effort available by default in your
client and adjust based on the task. Start with High for Luna or Light
for Astra.
In configuration, Astra’s Light setting is low. Increase the effort for tasks
that need more planning, analysis, or checking.
- Light in the ChatGPT desktop app, ChatGPT Work on the web, and IDE extension, or Low in the CLI, suits quick, well-scoped tasks.
- Medium balances speed and depth for tasks that need more planning.
- High and Extra High suit difficult work with multiple steps, sources, or tradeoffs.
Reasoning efforts don’t map exactly between model generations. Try a familiar task at a lower setting and adjust based on the result.
Know when to use Max or Ultra
Max gives the selected model more time to reason about a single task. Use it for the hardest problems, when depth matters more than speed or usage. If you don’t see Max in the ChatGPT desktop app, check your app settings.
Ultra uses subagents to handle separate parts of a complex task in parallel. Choose it when you can divide the work into meaningful parts. Most tasks do not need Max or Ultra.
GPT-6 Luna supports reasoning efforts up to Max, but not Ultra.
In the CLI, use /model, choose a model, then select More reasoning…
to see its supported Max or Ultra options.
If your model supports Ultra but it doesn’t appear in the desktop app’s model slider, go to Settings > Configuration, then turn on Ultra in model picker slider.
GPT-5.5 retirement
Effective October 14, 2026, at noon PDT (19:00 UTC), GPT-5.5 retires from Codex for users signed in with ChatGPT. The OpenAI API and Codex sessions authenticated with an API key aren’t affected.
For Codex with ChatGPT sign-in, choose an available replacement:
- On Plus, Pro, Business, Enterprise, and Edu plans, choose GPT-6.1 Sol
(
gpt-6.1-sol) when available. - On Free and Go plans, choose GPT-6 Luna (
gpt-6-luna) in the desktop app when available.
Replace gpt-5.5 in workspace defaults, saved model settings, managed
configurations, custom agents, scheduled tasks, and scripts that select a model.
See workspace model availability for administrator guidance.
Other models
With API-key authentication, model availability follows your API organization and project. GPT-5.5 remains available through the OpenAI API and in Codex sessions authenticated with an API key. See the API model catalog.
For a custom model provider or gateway, use a compatible Responses API endpoint.
See Custom model providers
for configuration and Gateway compatibility
for streaming, tool, and conversation requirements. Current Codex releases don’t
support wire_api = "chat" or a Chat Completions-only endpoint.
Deprecated Codex models
GPT-5.3-Codex-Spark retired on September 14, 2026, and is no longer available
in the ChatGPT desktop app, Codex CLI, or IDE extension. Update saved
configurations, custom agents, and scripts that select gpt-5.3-codex-spark
to use an available model. See the
Spark retirement notice.
GPT-5.5 retires from Codex with ChatGPT sign-in on October 14, 2026, at noon
PDT (19:00 UTC). Replace gpt-5.5 with a model available to your account and
client. API-key sessions aren’t affected. See
GPT-5.5 retirement for plan-specific
replacements and the migration checklist.
The gpt-5.4 and gpt-5.4-mini models retired from Codex with ChatGPT sign-in
on August 31, 2026. Replace gpt-5.4 with gpt-6-sol and
gpt-5.4-mini with gpt-6-luna when available to your plan and client.
In Enterprise and Edu, an administrator must enable Luna first. Update workspace
defaults, saved model settings, managed configurations, custom agents, and
scheduled tasks with an available replacement.
The gpt-5.2 and gpt-5.3-codex models are already deprecated in Codex when
you sign in with ChatGPT. Update scripts, configuration files, and
codex exec --model commands that still reference those models.
The OpenAI API and Codex authenticated with your own API key aren’t affected by the GPT-5.4 retirement. For current API model availability, see the API models page.
Configure your default local model
The ChatGPT desktop app, Codex CLI, and IDE extension use the same config.toml
configuration file. To specify a model, add a
model entry to your configuration file. If you don’t specify a model, the
ChatGPT desktop app, Codex CLI, or IDE extension uses a recommended model.
model = "gpt-6.1-sol"