檔案輸入 | OpenAI API

檔案輸入的支援範圍取決於 API 端點。Responses API 接受下列檔案類型作為 input_file 項目。Chat Completions 則僅接受 PDF 檔案作為 file 內容部分。

輸入方式	Responses API	Chat Completions
Base64 編碼的檔案資料（`file_data`）	下列支援的檔案類型	僅限 PDF
已上傳檔案的 ID（`file_id`）	下列支援的檔案類型	僅限 PDF
外部檔案 URL（`file_url`）	下列支援的檔案類型	不支援

若要輸入非 PDF 檔案，請使用 Responses API。若要在 Chat Completions 中使用檔案中的文字，請在應用程式中讀取檔案，並將其內容作為 text 內容部分傳送。

運作方式

在 Responses API 中，input_file 的處理方式取決於檔案類型：

PDF 檔案：使用具備視覺能力的模型（例如 gpt-4o 及後續模型）時，API 會同時擷取文字和頁面影像，並將兩者傳送給模型。
非 PDF 格式的文件與文字檔案 （例如 .docx、.pptx、.txt 和程式碼檔案）：API 僅擷取文字。
試算表檔案 （例如 .xlsx、.csv、.tsv）：API 會執行試算表專用的增強處理流程（詳見下文）。

如果以下相關工具更適合你的任務，請使用這些工具：

使用檔案搜尋檢索大型檔案，而非直接以 input_file 傳入整份檔案。
對於以試算表為主且需要詳細分析的任務，例如彙總、聯結、繪製圖表或自訂計算，請使用託管 Shell 環境。

非 PDF 檔案的影像與圖表限制

對於非 PDF 檔案，Responses API 不會擷取內嵌的影像或圖表作為模型的上下文。

若要保留圖表與示意圖的細節，請先將檔案轉換為 PDF，再以 input_file 傳送 PDF。

試算表增強處理的運作方式

對於試算表類型的檔案（例如 .xlsx、.xls、.csv、.tsv 和 .iif），Responses API 會使用試算表專用的增強處理流程。

API 不會將整張工作表傳給模型，而是最多解析每張工作表的前 1,000 列，並加入模型產生的摘要與標頭中繼資料，讓模型能以較精簡的結構化資料檢視進行處理。

PDF 細節層級

在 Responses API 中輸入 PDF 時，可以將 input_file 項目的選用欄位 detail 設為 auto、low 或 high，以控制 API 處理頁面影像的方式。若省略此欄位，detail 預設為 auto。對於 GPT-5.6 及後續模型，auto 會採用 high；較早的模型則採用 low。使用 low 可減少輸入 Token 用量；使用 high 則可保留更多視覺細節，例如密集的圖表、小字或示意圖。

detail 設定僅影響 PDF 頁面影像的處理方式。從 PDF 擷取的文字仍會納入。Chat Completions 的檔案輸入不支援 detail。

以下是明確指定高細節層級的最簡 Responses API 請求主體範例：

1234567891011121314151617181920{
  "model": "gpt-4.1",
  "input": [
    {
      "role": "user",
      "content": [
        {
          "type": "input_file",
          "filename": "document.pdf",
          "file_data": "data:application/pdf;base64,...",
          "detail": "high"
        },
        {
          "type": "input_text",
          "text": "Summarize this document."
        }
      ]
    }
  ]
}

支援的檔案類型

下表列出 Responses API 可接受作為 input_file 項目的常見檔案類型。完整的副檔名與 MIME 類型清單列於本頁後段。Chat Completions 的 file_data 與 file_id 都僅支援 .pdf（application/pdf）。

類別	常見副檔名
PDF 檔案	`.pdf`
文字與程式碼	`.txt`、`.md`、`.json`、`.html`、`.xml`、程式碼檔案
豐富格式文件	`.doc`、`.docx`、`.rtf`、`.odt`
簡報	`.ppt`、`.pptx`
試算表	`.csv`、`.xls`、`.xlsx`

檔案 URL

你可以透過外部 URL 連結提供檔案輸入。

使用外部檔案 URL

curl

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23import OpenAI from "openai";
const client = new OpenAI();

const response = await client.responses.create({
  model: "gpt-6-astra",
  input: [
    {
      role: "user",
      content: [
        {
          type: "input_text",
          text: "Analyze the letter and provide a summary of the key points.",
        },
        {
          type: "input_file",
          file_url: "https://www.berkshirehathaway.com/letters/2024ltr.pdf",
        },
      ],
    },
  ],
});

console.log(response.output_text);

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24from openai import OpenAI

client = OpenAI()

response = client.responses.create(
    model="gpt-6-astra",
    input=[
        {
            "role": "user",
            "content": [
                {
                    "type": "input_text",
                    "text": "Analyze the letter and provide a summary of the key points.",
                },
                {
                    "type": "input_file",
                    "file_url": "https://www.berkshirehathaway.com/letters/2024ltr.pdf",
                },
            ],
        },
    ],
)

print(response.output_text)

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41package main

import (
	"context"
	"fmt"

	"github.com/openai/openai-go/v3"
	"github.com/openai/openai-go/v3/responses"
)

func main() {
	client := openai.NewClient()

	response, err := client.Responses.New(context.Background(), responses.ResponseNewParams{
		Model: "gpt-6-astra",
		Input: responses.ResponseNewParamsInputUnion{
			OfInputItemList: responses.ResponseInputParam{
				responses.ResponseInputItemParamOfMessage(
					responses.ResponseInputMessageContentListParam{
						responses.ResponseInputContentParamOfInputText(
							"Analyze the letter and provide a summary of the key points.",
						),
						{
							OfInputFile: &responses.ResponseInputFileParam{
								FileURL: openai.String(
									"https://www.berkshirehathaway.com/letters/2024ltr.pdf",
								),
							},
						},
					},
					responses.EasyInputMessageRoleUser,
				),
			},
		},
	})
	if err != nil {
		panic(err)
	}

	fmt.Println(response.OutputText())
}

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30import com.openai.client.OpenAIClient;
import com.openai.client.okhttp.OpenAIOkHttpClient;
import com.openai.models.responses.ResponseCreateParams;
import com.openai.models.responses.ResponseInputFile;
import com.openai.models.responses.ResponseInputItem;
import java.util.List;

ResponseCreateParams params =
    ResponseCreateParams.builder()
        .model("gpt-6-astra")
        .inputOfResponse(
            List.of(
                ResponseInputItem.ofMessage(
                    ResponseInputItem.Message.builder()
                        .role(ResponseInputItem.Message.Role.USER)
                        .addInputTextContent(
                            "Analyze the letter and provide a summary of the key points.")
                        .addContent(
                            ResponseInputFile.builder()
                                .fileUrl(
                                    "https://www.berkshirehathaway.com/letters/2024ltr.pdf")
                                .build())
                        .build())))
        .build();

client.responses().create(params).output().stream()
    .flatMap(item -> item.message().stream())
    .flatMap(message -> message.content().stream())
    .flatMap(content -> content.outputText().stream())
    .forEach(text -> System.out.println(text.text()));

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25using OpenAI.Responses;
#pragma warning disable OPENAI001

string key = Environment.GetEnvironmentVariable("OPENAI_API_KEY")!;
ResponsesClient client = new(key);

Uri fileUrl = new(
    "https://www.berkshirehathaway.com/letters/2024ltr.pdf"
);

ResponseResult response = await client.CreateResponseAsync(
    "gpt-6-astra",
    [
        ResponseItem.CreateUserMessageItem(
            [
                ResponseContentPart.CreateInputTextPart(
                    "Analyze the letter and provide a summary of the key points."
                ),
                ResponseContentPart.CreateInputFilePart(fileUrl),
            ]
        ),
    ]
);

Console.WriteLine(response.GetOutputText());

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24require "openai"

openai = OpenAI::Client.new

response = openai.responses.create(
  model: "gpt-6-astra",
  input: [
    {
      role: "user",
      content: [
        {
          type: "input_text",
          text: "Analyze the letter and provide a summary of the key points."
        },
        {
          type: "input_file",
          file_url: "https://www.berkshirehathaway.com/letters/2024ltr.pdf"
        }
      ]
    }
  ]
)

puts(response.output_text)

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21curl "https://api.openai.com/v1/responses" \
    -H "Content-Type: application/json" \
    -H "Authorization: Bearer $OPENAI_API_KEY" \
    -d '{
        "model": "gpt-6-astra",
        "input": [
            {
                "role": "user",
                "content": [
                    {
                        "type": "input_text",
                        "text": "Analyze the letter and provide a summary of the key points."
                    },
                    {
                        "type": "input_file",
                        "file_url": "https://www.berkshirehathaway.com/letters/2024ltr.pdf"
                    }
                ]
            }
        ]
    }'

上傳檔案

以下範例使用 Files API 上傳檔案，再於傳送給模型的請求中參照該檔案的 ID。

上傳檔案

curl

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29import fs from "fs";
import OpenAI from "openai";
const client = new OpenAI();

const file = await client.files.create({
  file: fs.createReadStream("fixtures/draconomicon.pdf"),
  purpose: "user_data",
});

const response = await client.responses.create({
  model: "gpt-6-astra",
  input: [
    {
      role: "user",
      content: [
        {
          type: "input_file",
          file_id: file.id,
        },
        {
          type: "input_text",
          text: "What is the first dragon in the book?",
        },
      ],
    },
  ],
});

console.log(response.output_text);

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26from openai import OpenAI

client = OpenAI()

file = client.files.create(file=open("draconomicon.pdf", "rb"), purpose="user_data")

response = client.responses.create(
    model="gpt-6-astra",
    input=[
        {
            "role": "user",
            "content": [
                {
                    "type": "input_file",
                    "file_id": file.id,
                },
                {
                    "type": "input_text",
                    "text": "What is the first dragon in the book?",
                },
            ],
        }
    ],
)

print(response.output_text)

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54package main

import (
	"context"
	"fmt"
	"os"

	"github.com/openai/openai-go/v3"
	"github.com/openai/openai-go/v3/responses"
)

func main() {
	client := openai.NewClient()

	file, err := os.Open("draconomicon.pdf")
	if err != nil {
		panic(err)
	}
	defer file.Close()

	uploadedFile, err := client.Files.New(context.Background(), openai.FileNewParams{
		File:    file,
		Purpose: openai.FilePurposeUserData,
	})
	if err != nil {
		panic(err)
	}

	response, err := client.Responses.New(context.Background(), responses.ResponseNewParams{
		Model: "gpt-6-astra",
		Input: responses.ResponseNewParamsInputUnion{
			OfInputItemList: responses.ResponseInputParam{
				responses.ResponseInputItemParamOfMessage(
					responses.ResponseInputMessageContentListParam{
						{
							OfInputFile: &responses.ResponseInputFileParam{
								FileID: openai.String(uploadedFile.ID),
							},
						},
						responses.ResponseInputContentParamOfInputText(
							"What is the first dragon in the book?",
						),
					},
					responses.EasyInputMessageRoleUser,
				),
			},
		},
	})
	if err != nil {
		panic(err)
	}

	fmt.Println(response.OutputText())
}

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40import com.openai.client.OpenAIClient;
import com.openai.client.okhttp.OpenAIOkHttpClient;
import com.openai.models.files.FileCreateParams;
import com.openai.models.files.FilePurpose;
import com.openai.models.responses.ResponseCreateParams;
import com.openai.models.responses.ResponseInputFile;
import com.openai.models.responses.ResponseInputItem;
import java.nio.file.Path;
import java.util.List;

var file =
    client
        .files()
        .create(
            FileCreateParams.builder()
                .file(Path.of(System.getenv("OPENAI_EXAMPLE_FILE_PATH")))
                .purpose(FilePurpose.USER_DATA)
                .build());

var response =
    client
        .responses()
        .create(
            ResponseCreateParams.builder()
                .model("gpt-6-astra")
                .inputOfResponse(
                    List.of(
                        ResponseInputItem.ofMessage(
                            ResponseInputItem.Message.builder()
                                .role(ResponseInputItem.Message.Role.USER)
                                .addContent(
                                    ResponseInputFile.builder().fileId(file.id()).build())
                                .addInputTextContent("What is the first dragon in the book?")
                                .build())))
                .build());
response.output().stream()
    .flatMap(item -> item.message().stream())
    .flatMap(message -> message.content().stream())
    .flatMap(content -> content.outputText().stream())
    .forEach(text -> System.out.println(text.text()));

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29using OpenAI.Files;
using OpenAI.Responses;
#pragma warning disable OPENAI001

string key = Environment.GetEnvironmentVariable("OPENAI_API_KEY")!;
ResponsesClient client = new(key);

OpenAIFileClient files = new(key);

OpenAIFile file = await files.UploadFileAsync(
    "draconomicon.pdf",
    FileUploadPurpose.UserData
);

ResponseResult response = await client.CreateResponseAsync(
    "gpt-6-astra",
    [
        ResponseItem.CreateUserMessageItem(
            [
                ResponseContentPart.CreateInputFilePart(file.Id),
                ResponseContentPart.CreateInputTextPart(
                    "What is the first dragon in the book?"
                ),
            ]
        ),
    ]
);

Console.WriteLine(response.GetOutputText());

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30require "openai"
require "pathname"

openai = OpenAI::Client.new

file = openai.files.create(
  file: Pathname("draconomicon.pdf"),
  purpose: "user_data"
)

response = openai.responses.create(
  model: "gpt-6-astra",
  input: [
    {
      role: "user",
      content: [
        {
          type: "input_file",
          file_id: file.id
        },
        {
          type: "input_text",
          text: "What is the first dragon in the book?"
        }
      ]
    }
  ]
)

puts(response.output_text)

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26curl https://api.openai.com/v1/files \
    -H "Authorization: Bearer $OPENAI_API_KEY" \
    -F purpose="user_data" \
    -F file="@draconomicon.pdf"

curl "https://api.openai.com/v1/responses" \
    -H "Content-Type: application/json" \
    -H "Authorization: Bearer $OPENAI_API_KEY" \
    -d '{
        "model": "gpt-6-astra",
        "input": [
            {
                "role": "user",
                "content": [
                    {
                        "type": "input_file",
                        "file_id": "file-6F2ksmvXxt4VdoqmHRw6kL"
                    },
                    {
                        "type": "input_text",
                        "text": "What is the first dragon in the book?"
                    }
                ]
            }
        ]
    }'

上傳檔案

curl

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31import fs from "fs";
import OpenAI from "openai";
const client = new OpenAI();

const file = await client.files.create({
  file: fs.createReadStream("fixtures/draconomicon.pdf"),
  purpose: "user_data",
});

const completion = await client.chat.completions.create({
  model: "gpt-6-astra",
  messages: [
    {
      role: "user",
      content: [
        {
          type: "file",
          file: {
            file_id: file.id,
          },
        },
        {
          type: "text",
          text: "What is the first dragon in the book?",
        },
      ],
    },
  ],
});

console.log(completion.choices[0].message.content);

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28from openai import OpenAI

client = OpenAI()

file = client.files.create(file=open("draconomicon.pdf", "rb"), purpose="user_data")

completion = client.chat.completions.create(
    model="gpt-6-astra",
    messages=[
        {
            "role": "user",
            "content": [
                {
                    "type": "file",
                    "file": {
                        "file_id": file.id,
                    },
                },
                {
                    "type": "text",
                    "text": "What is the first dragon in the book?",
                },
            ],
        }
    ],
)

print(completion.choices[0].message.content)

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44package main

import (
	"context"
	"fmt"
	"os"

	"github.com/openai/openai-go/v3"
)

func main() {
	client := openai.NewClient()

	file, err := os.Open("draconomicon.pdf")
	if err != nil {
		panic(err)
	}
	defer file.Close()

	uploadedFile, err := client.Files.New(context.Background(), openai.FileNewParams{
		File:    file,
		Purpose: openai.FilePurposeUserData,
	})
	if err != nil {
		panic(err)
	}

	completion, err := client.Chat.Completions.New(context.Background(), openai.ChatCompletionNewParams{
		Model: "gpt-6-astra",
		Messages: []openai.ChatCompletionMessageParamUnion{
			openai.UserMessage([]openai.ChatCompletionContentPartUnionParam{
				openai.FileContentPart(openai.ChatCompletionContentPartFileFileParam{
					FileID: openai.String(uploadedFile.ID),
				}),
				openai.TextContentPart("What is the first dragon in the book?"),
			}),
		},
	})
	if err != nil {
		panic(err)
	}

	fmt.Println(completion.Choices[0].Message.Content)
}

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43import com.openai.client.OpenAIClient;
import com.openai.client.okhttp.OpenAIOkHttpClient;
import com.openai.models.chat.completions.ChatCompletionContentPart;
import com.openai.models.chat.completions.ChatCompletionContentPartText;
import com.openai.models.chat.completions.ChatCompletionCreateParams;
import com.openai.models.chat.completions.ChatCompletionUserMessageParam;
import com.openai.models.files.FileCreateParams;
import com.openai.models.files.FilePurpose;
import java.nio.file.Path;
import java.util.List;

var file =
    client
        .files()
        .create(
            FileCreateParams.builder()
                .file(Path.of(System.getenv("OPENAI_EXAMPLE_FILE_PATH")))
                .purpose(FilePurpose.USER_DATA)
                .build());

var filePart =
    ChatCompletionContentPart.ofFile(
        ChatCompletionContentPart.File.builder()
            .file(ChatCompletionContentPart.File.FileObject.builder().fileId(file.id()).build())
            .build());
var textPart =
    ChatCompletionContentPart.ofText(
        ChatCompletionContentPartText.builder().text("Summarize this PDF.").build());
var completion =
    client
        .chat()
        .completions()
        .create(
            ChatCompletionCreateParams.builder()
                .model("gpt-6-astra")
                .addMessage(
                    ChatCompletionUserMessageParam.builder()
                        .contentOfArrayOfContentParts(List.of(filePart, textPart))
                        .build())
                .build());
completion.choices().stream()
    .flatMap(choice -> choice.message().content().stream())
    .forEach(System.out::println);

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22using OpenAI.Chat;
using OpenAI.Files;
#pragma warning disable OPENAI001

string key = Environment.GetEnvironmentVariable("OPENAI_API_KEY")!;
string model = "gpt-6-astra";
ChatClient client = new(model, key);

OpenAIFileClient files = new(key);

OpenAIFile file = await files.UploadFileAsync("draconomicon.pdf", FileUploadPurpose.UserData);

ChatCompletion completion = await client.CompleteChatAsync(
    new UserChatMessage(
        [
            ChatMessageContentPart.CreateFilePart(file.Id),
            ChatMessageContentPart.CreateTextPart("What is the first dragon in the book?"),
        ]
    )
);

Console.WriteLine(completion.Content[0].Text);

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28require "openai"
require "pathname"

client = OpenAI::Client.new
file = client.files.create(
  file: Pathname("draconomicon.pdf"),
  purpose: :user_data
)
completion = client.chat.completions.create(
  model: "gpt-6-astra",
  messages: [
    {
      role: :user,
      content: [
        {
          type: :file,
          file: { file_id: file.id }
        },
        {
          type: :text,
          text: "Summarize this PDF."
        }
      ]
    }
  ]
)

puts(completion.choices.fetch(0).message.content)

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28curl https://api.openai.com/v1/files \
    -H "Authorization: Bearer $OPENAI_API_KEY" \
    -F purpose="user_data" \
    -F file="@draconomicon.pdf"

curl "https://api.openai.com/v1/chat/completions" \
    -H "Content-Type: application/json" \
    -H "Authorization: Bearer $OPENAI_API_KEY" \
    -d '{
        "model": "gpt-6-astra",
        "messages": [
            {
                "role": "user",
                "content": [
                    {
                        "type": "file",
                        "file": {
                            "file_id": "file-6F2ksmvXxt4VdoqmHRw6kL"
                        }
                    },
                    {
                        "type": "text",
                        "text": "What is the first dragon in the book?"
                    }
                ]
            }
        ]
    }'

Base64 編碼的檔案

你也可以使用 Base64 編碼的檔案資料傳送檔案輸入。

傳送 Base64 編碼的檔案

curl

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28import fs from "fs";
import OpenAI from "openai";
const client = new OpenAI();

const data = fs.readFileSync("fixtures/draconomicon.pdf");
const base64String = data.toString("base64");

const response = await client.responses.create({
  model: "gpt-6-astra",
  input: [
    {
      role: "user",
      content: [
        {
          type: "input_file",
          filename: "draconomicon.pdf",
          file_data: `data:application/pdf;base64,${base64String}`,
        },
        {
          type: "input_text",
          text: "What is the first dragon in the book?",
        },
      ],
    },
  ],
});

console.log(response.output_text);

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31import base64
from openai import OpenAI

client = OpenAI()

with open("draconomicon.pdf", "rb") as f:
    data = f.read()

base64_string = base64.b64encode(data).decode("utf-8")

response = client.responses.create(
    model="gpt-6-astra",
    input=[
        {
            "role": "user",
            "content": [
                {
                    "type": "input_file",
                    "filename": "draconomicon.pdf",
                    "file_data": f"data:application/pdf;base64,{base64_string}",
                },
                {
                    "type": "input_text",
                    "text": "What is the first dragon in the book?",
                },
            ],
        },
    ],
)

print(response.output_text)

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48package main

import (
	"context"
	"encoding/base64"
	"fmt"
	"os"

	"github.com/openai/openai-go/v3"
	"github.com/openai/openai-go/v3/responses"
)

func main() {
	client := openai.NewClient()

	data, err := os.ReadFile("draconomicon.pdf")
	if err != nil {
		panic(err)
	}
	fileData := "data:application/pdf;base64," + base64.StdEncoding.EncodeToString(data)

	response, err := client.Responses.New(context.Background(), responses.ResponseNewParams{
		Model: "gpt-6-astra",
		Input: responses.ResponseNewParamsInputUnion{
			OfInputItemList: responses.ResponseInputParam{
				responses.ResponseInputItemParamOfMessage(
					responses.ResponseInputMessageContentListParam{
						{
							OfInputFile: &responses.ResponseInputFileParam{
								Filename: openai.String("draconomicon.pdf"),
								FileData: openai.String(fileData),
							},
						},
						responses.ResponseInputContentParamOfInputText(
							"What is the first dragon in the book?",
						),
					},
					responses.EasyInputMessageRoleUser,
				),
			},
		},
	})
	if err != nil {
		panic(err)
	}

	fmt.Println(response.OutputText())
}

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36import com.openai.client.OpenAIClient;
import com.openai.client.okhttp.OpenAIOkHttpClient;
import com.openai.models.responses.ResponseCreateParams;
import com.openai.models.responses.ResponseInputFile;
import com.openai.models.responses.ResponseInputItem;
import java.io.IOException;
import java.nio.file.Files;
import java.nio.file.Path;
import java.util.Base64;
import java.util.List;

String pdfData =
    Base64.getEncoder()
        .encodeToString(Files.readAllBytes(Path.of(System.getenv("OPENAI_EXAMPLE_FILE_PATH"))));
ResponseCreateParams params =
    ResponseCreateParams.builder()
        .model("gpt-6-astra")
        .inputOfResponse(
            List.of(
                ResponseInputItem.ofMessage(
                    ResponseInputItem.Message.builder()
                        .role(ResponseInputItem.Message.Role.USER)
                        .addContent(
                            ResponseInputFile.builder()
                                .filename("document.pdf")
                                .fileData("data:application/pdf;base64," + pdfData)
                                .build())
                        .addInputTextContent("Summarize this document.")
                        .build())))
        .build();

client.responses().create(params).output().stream()
    .flatMap(item -> item.message().stream())
    .flatMap(message -> message.content().stream())
    .flatMap(content -> content.outputText().stream())
    .forEach(text -> System.out.println(text.text()));

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26using OpenAI.Responses;
#pragma warning disable OPENAI001

string key = Environment.GetEnvironmentVariable("OPENAI_API_KEY")!;
ResponsesClient client = new(key);

BinaryData fileBytes = BinaryData.FromBytes(await File.ReadAllBytesAsync("draconomicon.pdf"));
ResponseResult response = await client.CreateResponseAsync(
    "gpt-6-astra",
    [
        ResponseItem.CreateUserMessageItem(
            [
                ResponseContentPart.CreateInputFilePart(
                    fileBytes,
                    "application/pdf",
                    "draconomicon.pdf"
                ),
                ResponseContentPart.CreateInputTextPart(
                    "What is the first dragon in the book?"
                ),
            ]
        ),
    ]
);

Console.WriteLine(response.GetOutputText());

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26require "base64"
require "openai"

client = OpenAI::Client.new
pdf_data = Base64.strict_encode64(File.binread("draconomicon.pdf"))
response = client.responses.create(
  model: "gpt-6-astra",
  input: [
    {
      role: :user,
      content: [
        {
          type: :input_file,
          filename: "document.pdf",
          file_data: "data:application/pdf;base64,#{pdf_data}"
        },
        {
          type: :input_text,
          text: "Summarize this document."
        }
      ]
    }
  ]
)

puts(response.output_text)

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22curl "https://api.openai.com/v1/responses" \
    -H "Content-Type: application/json" \
    -H "Authorization: Bearer $OPENAI_API_KEY" \
    -d '{
        "model": "gpt-6-astra",
        "input": [
            {
                "role": "user",
                "content": [
                    {
                        "type": "input_file",
                        "filename": "draconomicon.pdf",
                        "file_data": "...base64 encoded PDF bytes here..."
                    },
                    {
                        "type": "input_text",
                        "text": "What is the first dragon in the book?"
                    }
                ]
            }
        ]
    }'

傳送 Base64 編碼的檔案

curl

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30import fs from "fs";
import OpenAI from "openai";
const client = new OpenAI();

const data = fs.readFileSync("fixtures/draconomicon.pdf");
const base64String = data.toString("base64");

const completion = await client.chat.completions.create({
  model: "gpt-6-astra",
  messages: [
    {
      role: "user",
      content: [
        {
          type: "file",
          file: {
            filename: "draconomicon.pdf",
            file_data: `data:application/pdf;base64,${base64String}`,
          },
        },
        {
          type: "text",
          text: "What is the first dragon in the book?",
        },
      ],
    },
  ],
});

console.log(completion.choices[0].message.content);

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33import base64
from openai import OpenAI

client = OpenAI()

with open("draconomicon.pdf", "rb") as f:
    data = f.read()

base64_string = base64.b64encode(data).decode("utf-8")

completion = client.chat.completions.create(
    model="gpt-6-astra",
    messages=[
        {
            "role": "user",
            "content": [
                {
                    "type": "file",
                    "file": {
                        "filename": "draconomicon.pdf",
                        "file_data": f"data:application/pdf;base64,{base64_string}",
                    },
                },
                {
                    "type": "text",
                    "text": "What is the first dragon in the book?",
                },
            ],
        },
    ],
)

print(completion.choices[0].message.content)

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38package main

import (
	"context"
	"encoding/base64"
	"fmt"
	"os"

	"github.com/openai/openai-go/v3"
)

func main() {
	client := openai.NewClient()

	data, err := os.ReadFile("draconomicon.pdf")
	if err != nil {
		panic(err)
	}
	fileData := "data:application/pdf;base64," + base64.StdEncoding.EncodeToString(data)

	completion, err := client.Chat.Completions.New(context.Background(), openai.ChatCompletionNewParams{
		Model: "gpt-6-astra",
		Messages: []openai.ChatCompletionMessageParamUnion{
			openai.UserMessage([]openai.ChatCompletionContentPartUnionParam{
				openai.FileContentPart(openai.ChatCompletionContentPartFileFileParam{
					Filename: openai.String("draconomicon.pdf"),
					FileData: openai.String(fileData),
				}),
				openai.TextContentPart("What is the first dragon in the book?"),
			}),
		},
	})
	if err != nil {
		panic(err)
	}

	fmt.Println(completion.Choices[0].Message.Content)
}

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40import com.openai.client.OpenAIClient;
import com.openai.client.okhttp.OpenAIOkHttpClient;
import com.openai.models.chat.completions.ChatCompletionContentPart;
import com.openai.models.chat.completions.ChatCompletionContentPartText;
import com.openai.models.chat.completions.ChatCompletionCreateParams;
import com.openai.models.chat.completions.ChatCompletionUserMessageParam;
import java.io.IOException;
import java.nio.file.Files;
import java.nio.file.Path;
import java.util.Base64;
import java.util.List;

String encoded =
    Base64.getEncoder()
        .encodeToString(Files.readAllBytes(Path.of(System.getenv("OPENAI_EXAMPLE_FILE_PATH"))));
ChatCompletionContentPart file =
    ChatCompletionContentPart.ofFile(
        ChatCompletionContentPart.File.builder()
            .file(
                ChatCompletionContentPart.File.FileObject.builder()
                    .filename("document.pdf")
                    .fileData("data:application/pdf;base64," + encoded)
                    .build())
            .build());
ChatCompletionContentPart text =
    ChatCompletionContentPart.ofText(
        ChatCompletionContentPartText.builder().text("Summarize this document.").build());

ChatCompletionCreateParams params =
    ChatCompletionCreateParams.builder()
        .model("gpt-6-astra")
        .addMessage(
            ChatCompletionUserMessageParam.builder()
                .contentOfArrayOfContentParts(List.of(file, text))
                .build())
        .build();

client.chat().completions().create(params).choices().stream()
    .flatMap(choice -> choice.message().content().stream())
    .forEach(System.out::println);

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22using OpenAI.Chat;
#pragma warning disable OPENAI001

string key = Environment.GetEnvironmentVariable("OPENAI_API_KEY")!;
string model = "gpt-6-astra";
ChatClient client = new(model, key);

BinaryData fileBytes = BinaryData.FromBytes(await File.ReadAllBytesAsync("draconomicon.pdf"));
ChatCompletion completion = await client.CompleteChatAsync(
    new UserChatMessage(
        [
            ChatMessageContentPart.CreateFilePart(
                fileBytes,
                "application/pdf",
                "draconomicon.pdf"
            ),
            ChatMessageContentPart.CreateTextPart("What is the first dragon in the book?"),
        ]
    )
);

Console.WriteLine(completion.Content[0].Text);

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28require "base64"
require "openai"

client = OpenAI::Client.new
pdf_data = Base64.strict_encode64(File.binread("draconomicon.pdf"))
completion = client.chat.completions.create(
  model: "gpt-6-astra",
  messages: [
    {
      role: :user,
      content: [
        {
          type: :file,
          file: {
            filename: "document.pdf",
            file_data: "data:application/pdf;base64,#{pdf_data}"
          }
        },
        {
          type: :text,
          text: "Summarize this document."
        }
      ]
    }
  ]
)

puts(completion.choices.fetch(0).message.content)

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24curl "https://api.openai.com/v1/chat/completions" \
    -H "Content-Type: application/json" \
    -H "Authorization: Bearer $OPENAI_API_KEY" \
    -d '{
        "model": "gpt-6-astra",
        "messages": [
            {
                "role": "user",
                "content": [
                    {
                        "type": "file",
                        "file": {
                            "filename": "draconomicon.pdf",
                            "file_data": "...base64 encoded bytes here..."
                        }
                    },
                    {
                        "type": "text",
                        "text": "What is the first dragon in the book?"
                    }
                ]
            }
        ]
    }'

使用注意事項

使用檔案輸入時，請留意以下限制：

Token 用量： 解析 PDF 時，擷取的文字和頁面影像都會納入上下文，因此可能增加 Token 用量。在 Responses API 中，將 detail 設為 auto（預設值）、low 或 high，可控制 PDF 頁面影像的視覺細節量。大規模部署前，請先瞭解定價及對 Token 用量的影響。瞭解定價詳情。
檔案大小限制： 單一請求可以包含多個檔案，但每個檔案都必須小於 50 MB。請求中所有檔案的總大小上限為 50 MB。
支援的模型： 若要在解析 PDF 時同時處理文字和頁面影像，需要使用具備視覺能力的模型，例如 gpt-4o 及後續模型。
檔案上傳用途： 你可以指定任何支援的用途來上傳檔案，但對於打算作為模型輸入的檔案，請使用 user_data。

支援的檔案類型完整清單

此清單適用於 Responses API。Chat Completions 的 file_data 與 file_id 都僅支援 .pdf （application/pdf）。

類別	副檔名	MIME 類型
PDF 檔案	PDF 檔案（`.pdf`）	`application/pdf`
試算表	Excel 工作表（`.xla`, `.xlb`, `.xlc`, `.xlm`, `.xls`, `.xlsx`, `.xlt`, `.xlw`）	`application/vnd.openxmlformats-officedocument.spreadsheetml.sheet`, `application/vnd.ms-excel`
試算表	CSV / TSV / IIF（`.csv`, `.tsv`, `.iif`）、Google Sheets	`text/csv`, `application/csv`, `text/tsv`, `text/x-iif`, `application/x-iif`, `application/vnd.google-apps.spreadsheet`
格式化文件	Word/ODT/RTF 文件（`.doc`, `.docx`, `.dot`, `.odt`, `.rtf`）、Pages、Google Docs	`application/vnd.openxmlformats-officedocument.wordprocessingml.document`, `application/msword`, `application/rtf`, `text/rtf`, `application/vnd.oasis.opendocument.text`, `application/vnd.apple.pages`, `application/vnd.google-apps.document`, `application/vnd.apple.iwork`
簡報	PowerPoint 投影片（`.pot`, `.ppa`, `.pps`, `.ppt`, `.pptx`, `.pwz`, `.wiz`）、Keynote、Google Slides	`application/vnd.openxmlformats-officedocument.presentationml.presentation`, `application/vnd.ms-powerpoint`, `application/vnd.apple.keynote`, `application/vnd.google-apps.presentation`, `application/vnd.apple.iwork`
文字與程式碼	文字／程式碼格式（`.asm`, `.bat`, `.c`, `.cc`, `.conf`, `.cpp`, `.css`, `.cxx`, `.def`, `.dic`, `.eml`, `.h`, `.hh`, `.htm`, `.html`, `.ics`, `.ifb`, `.in`, `.js`, `.json`, `.ksh`, `.list`, `.log`, `.markdown`, `.md`, `.mht`, `.mhtml`, `.mime`, `.mjs`, `.nws`, `.pl`, `.py`, `.rst`, `.s`, `.sql`, `.srt`, `.text`, `.txt`, `.vcf`, `.vtt`, `.xml`）	`application/javascript`, `application/typescript`, `text/xml`, `text/x-shellscript`, `text/x-rst`, `text/x-makefile`, `text/x-lisp`, `text/x-asm`, `text/vbscript`, `text/css`, `message/rfc822`, `application/x-sql`, `application/x-scala`, `application/x-rust`, `application/x-powershell`, `text/x-diff`, `text/x-patch`, `application/x-patch`, `text/plain`, `text/markdown`, `text/x-java`, `text/x-script.python`, `text/x-python`, `text/x-c`, `text/x-c++`, `text/x-golang`, `text/html`, `text/x-php`, `application/x-php`, `application/x-httpd-php`, `application/x-httpd-php-source`, `text/x-ruby`, `text/x-sh`, `text/x-bash`, `application/x-bash`, `text/x-zsh`, `text/x-tex`, `text/x-csharp`, `application/json`, `text/x-typescript`, `text/javascript`, `text/x-go`, `text/x-rust`, `text/x-scala`, `text/x-kotlin`, `text/x-swift`, `text/x-lua`, `text/x-r`, `text/x-R`, `text/x-julia`, `text/x-perl`, `text/x-objectivec`, `text/x-objectivec++`, `text/x-erlang`, `text/x-elixir`, `text/x-haskell`, `text/x-clojure`, `text/x-groovy`, `text/x-dart`, `text/x-awk`, `application/x-awk`, `text/jsx`, `text/tsx`, `text/x-handlebars`, `text/x-mustache`, `text/x-ejs`, `text/x-jinja2`, `text/x-liquid`, `text/x-erb`, `text/x-twig`, `text/x-pug`, `text/x-jade`, `text/x-tmpl`, `text/x-cmake`, `text/x-dockerfile`, `text/x-gradle`, `text/x-ini`, `text/x-properties`, `text/x-protobuf`, `application/x-protobuf`, `text/x-sql`, `text/x-sass`, `text/x-scss`, `text/x-less`, `text/x-hcl`, `text/x-terraform`, `application/x-terraform`, `text/x-toml`, `application/x-toml`, `application/graphql`, `application/x-graphql`, `text/x-graphql`, `application/x-ndjson`, `application/json5`, `application/x-json5`, `text/x-yaml`, `application/toml`, `application/x-yaml`, `application/yaml`, `text/x-astro`, `text/srt`, `application/x-subrip`, `text/x-subrip`, `text/vtt`, `text/x-vcard`, `text/calendar`

後續步驟

接下來，你可以參考以下資源：

在 Playground 中試用檔案輸入

使用 Playground 設計並反覆調整搭配檔案輸入的提示詞。

完整 API 參考文件

查看 API 參考文件，瞭解更多選項。

使用檔案搜尋處理大型語料庫

需要可擴展的搜尋功能時，請對分塊後的檔案進行檢索，而不是將整份檔案傳送到單一上下文視窗。

使用託管 Shell 環境深入分析試算表

使用託管 Shell 環境執行進階試算表工作流程，例如聯結、彙總及繪製圖表。

建議

開始使用

核心概念

SDK 與 CLI

資源

舊版 API

選擇模型

文字與程式碼

提示詞

推理

圖像與影片

即時互動與音訊

專用模型

智慧體 API

智慧體 SDK

ChatKit

搜尋與檢索

連接工具與資料

建立工具工作流程

電腦與程式碼

媒體

GPT-Live

Realtime API

開發語音應用

連線

音訊處理

正式上線

效能與品質

成本與吞吐量

安全與治理

基礎架構與存取

營運

核心概念

Plan

Build

Test and publish

Conversion specs

指南

資源

開始使用

指南

File Upload

API

Measurement

Advertiser API

API 參考文件

開始使用

基礎

探索

適用平台

版本資訊

工作流程

能力

參考資料

自訂

設定檔

智慧體組態

擴充 ChatGPT 與 Codex 的功能

Linux

Windows

開發工作流程

擴充與自動化

環境

使用 Codex 進行建置

第三方整合

參考資料

權限

Codex Security

網路安全

開始使用

ChatGPT Work

身分識別與身分驗證

工作區存取權、政策與模型

外掛程式與連接器控制措施

使用情況、治理與合規

部署與模型供應商

社群

部落格

社群

部落格