> ## Documentation Index
> Fetch the complete documentation index at: https://evolink.ai/docs/llms.txt
> Use this file to discover all available pages before exploring further.
# MiniMax-M3.1-Flash - Responses 接口
> - 使用 OpenAI Responses 协议调用 MiniMax-M3.1-Flash 模型
- **输入形态**：`input` 可以是一段纯文本，也可以是输入项数组（消息、函数调用、函数结果）
- **多模态输入**：在消息的 `content` 中混合 `input_text`、`input_image`、`input_video` 内容块
- **系统指令**：通过 `instructions` 下达，等价于首条 system 消息
- **思考模式**：始终开启，`reasoning.effort` 调节思考深度；思考内容以 `reasoning` 输出项返回
- **流式输出**：语义化 SSE 事件，以 `response.completed` 结束，**不发送 `[DONE]`**
- **工具调用**：支持 Function Calling
- **上下文缓存**：相同前缀自动命中，命中部分体现在 `usage.input_tokens_details.cached_tokens`
**无状态**：不保存响应，不支持 `previous_response_id`。多轮对话时，将上一轮 `output` 中的所有项按顺序原样追加到 `input`，保留 `reasoning` 项中的 `content`、`summary` 和 `encrypted_content` 等字段。执行工具后，再追加与函数调用项 `call_id` 相同的 `function_call_output` 项。
**流式事件**：`response.created`、`response.in_progress`、`response.output_item.added`、`response.content_part.added`、`response.reasoning_text.delta` / `response.reasoning_text.done`、`response.output_text.delta` / `response.output_text.done`、`response.content_part.done`、`response.output_item.done`，终止事件为 `response.completed`。每个事件都带 `sequence_number` 用于排序。
**BaseURL 说明**：默认 BaseURL 为 `https://direct.evolink.ai`，对文本模型支持更好，支持长连接；`https://api.evolink.ai` 是多模态主力地址，对文本模型作为备用地址使用。
## OpenAPI
````yaml cn/api-manual/language-series/minimax-m3.1-flash/minimax-m3.1-flash-responses.json POST /v1/responses
openapi: 3.1.0
info:
title: MiniMax-M3.1-Flash 完整参数文档 (Responses API)
description: |-
MiniMax-M3.1-Flash Responses 接口的完整 API 参考。
**模型能力**：
- 上下文长度：**1,000,000 tokens**（1M）
- 最大输出：**524,288 tokens**（512K），推荐 **131,072 tokens**（128K），思考 token 计入其中
- 思考模式：始终开启，通过 `reasoning.effort` 调节思考深度，默认 `max`
- 多模态输入：支持文本、图像、视频
- 工具调用：支持 Function Calling
- 上下文缓存：相同前缀自动命中，命中部分按缓存命中价计费
**无状态**：不保存响应，多轮对话需要客户端每次携带完整历史。
license:
name: MIT
version: 1.0.0
servers:
- url: https://direct.evolink.ai
description: 生产环境（推荐）
- url: https://api.evolink.ai
description: 备用地址
security:
- bearerAuth: []
tags:
- name: 对话生成
description: AI 对话生成相关接口
paths:
/v1/responses:
post:
tags:
- 对话生成
summary: MiniMax-M3.1-Flash Responses 接口（OpenAI 兼容）
description: >-
- 使用 OpenAI Responses 协议调用 MiniMax-M3.1-Flash 模型
- **输入形态**：`input` 可以是一段纯文本，也可以是输入项数组（消息、函数调用、函数结果）
- **多模态输入**：在消息的 `content` 中混合 `input_text`、`input_image`、`input_video`
内容块
- **系统指令**：通过 `instructions` 下达，等价于首条 system 消息
- **思考模式**：始终开启，`reasoning.effort` 调节思考深度；思考内容以 `reasoning` 输出项返回
- **流式输出**：语义化 SSE 事件，以 `response.completed` 结束，**不发送 `[DONE]`**
- **工具调用**：支持 Function Calling
- **上下文缓存**：相同前缀自动命中，命中部分体现在 `usage.input_tokens_details.cached_tokens`
**无状态**：不保存响应，不支持 `previous_response_id`。多轮对话时，将上一轮 `output`
中的所有项按顺序原样追加到 `input`，保留 `reasoning` 项中的 `content`、`summary` 和
`encrypted_content` 等字段。执行工具后，再追加与函数调用项 `call_id` 相同的
`function_call_output` 项。
**流式事件**：`response.created`、`response.in_progress`、`response.output_item.added`、`response.content_part.added`、`response.reasoning_text.delta`
/ `response.reasoning_text.done`、`response.output_text.delta` /
`response.output_text.done`、`response.content_part.done`、`response.output_item.done`，终止事件为
`response.completed`。每个事件都带 `sequence_number` 用于排序。
operationId: createResponseMiniMaxM31Flash
requestBody:
required: true
content:
application/json:
schema:
$ref: '#/components/schemas/ResponsesRequest'
examples:
simple_text:
summary: 单轮文本对话
value:
model: MiniMax-M3.1-Flash
input: Explain recursion in one sentence.
max_output_tokens: 131072
with_instructions:
summary: 使用系统指令
value:
model: MiniMax-M3.1-Flash
instructions: >-
You are a technical documentation assistant. Keep your
answers concise.
input: Explain how context caching works.
max_output_tokens: 131072
multi_turn:
summary: 多轮对话（输入项数组）
value:
model: MiniMax-M3.1-Flash
input:
- type: message
role: user
content: Suggest three names for a note-taking app.
- type: message
role: assistant
content: NoteFlow, ClearPage, QuickNote.
- type: message
role: user
content: Suggest five alternatives to the second name.
max_output_tokens: 131072
reasoning_effort:
summary: 调节思考深度
value:
model: MiniMax-M3.1-Flash
input: >-
A bat and a ball cost 1.10 in total. The bat costs 1.00 more
than the ball. How much does the ball cost?
reasoning:
effort: high
max_output_tokens: 131072
vision_url:
summary: 图像理解
value:
model: MiniMax-M3.1-Flash
input:
- type: message
role: user
content:
- type: input_text
text: Describe this image.
- type: input_image
image_url: https://example.com/photo.jpg
max_output_tokens: 131072
video_url:
summary: 视频理解
value:
model: MiniMax-M3.1-Flash
input:
- type: message
role: user
content:
- type: input_text
text: Summarize this video.
- type: input_video
video_url:
url: https://example.com/video.mp4
fps: 1
detail: default
max_output_tokens: 131072
function_calling:
summary: Function Calling 工具调用
value:
model: MiniMax-M3.1-Flash
input: Check the stock for product demo-item-001.
tools:
- type: function
name: get_product_stock
description: Get the current stock level for a product.
parameters:
type: object
properties:
product_id:
type: string
description: Product identifier.
required:
- product_id
max_output_tokens: 131072
function_result:
summary: 回传工具执行结果并续答
value:
model: MiniMax-M3.1-Flash
reasoning:
effort: medium
max_output_tokens: 131072
input:
- type: message
role: user
content: >-
Use get_product_stock to check product demo-item-001.
After receiving the tool result, report whether it is in
stock and its quantity. Do not guess the stock level.
- id: 0725233e825de1d0656437967fc024e1_msg
type: message
status: completed
role: assistant
content:
- type: output_text
text: I'll check the stock level for that product now.
annotations: []
- id: 0725233e825de1d0656437967fc024e1_fc_0
type: function_call
status: completed
call_id: call_function_izs7qcnzfksd_1
name: get_product_stock
arguments: '{"product_id": "demo-item-001"}'
- type: function_call_output
call_id: call_function_izs7qcnzfksd_1
output: >-
{"product_id": "demo-item-001", "in_stock": true,
"quantity": 24}
tools:
- type: function
name: get_product_stock
description: Get the current stock level for a product.
parameters:
type: object
properties:
product_id:
type: string
description: Product identifier.
required:
- product_id
tool_choice: auto
streaming:
summary: 流式输出
value:
model: MiniMax-M3.1-Flash
input: Write a short poem about autumn.
stream: true
max_output_tokens: 131072
reasoning_history:
summary: 携带完整思考历史继续对话
value:
model: MiniMax-M3.1-Flash
reasoning:
effort: medium
max_output_tokens: 131072
input:
- type: message
role: user
content: >-
Find the smallest positive integer n such that n mod 7 =
3, n mod 11 = 5, and n mod 13 = 7. Return only n.
- id: 0725233a85e23431681447327cbcd9d2_rs
type: reasoning
status: completed
summary: []
content:
- type: reasoning_text
text: >-
We need solve CRT. Need only n. Compute. n=3 mod7, 5
mod11,7 mod13. Product 1001. Let's solve.
Use n=3+7a. mod11: 3+7a=5 =>7a=2 mod11. inverse7=8
(56=1), a=16=5 mod11. a=5+11b. n=38+77b. mod13:
38+77b = 12+12b mod13 (77=12,38=12) equals7
=>12b=-5=8 mod13. inverse12=12, b=96 mod13 =5? 91
remainder5. n=38+385=423. Check 423 mod7 3 (420),
mod11 5 (418), mod13 7 (416). Smallest positive 423.
Only n.
encrypted_content: >-
Ahlub3JtYWwtcmVhc29uaW5nLTIwMjYxMDExjehfvIf2rUinmic9ZiAutly0DD3lHqWEni9ENDziUw0nX-F6VPZLM7x1NbRWbytJNcLH9XQIa-KLYgd-it6XmXzYLvck3xOPOG9wtDdxUnjdBhuuj9dkWrkx0kbK8MFT5UreZ7EcxoRZybrs3oELf2BK_WtDUFOm_FbZXKwNbg76pTCFDoRke0Er0GMWjmi5L8GE7hGnvvBQjHXrwQ88G0L2p6STay-n1FJ5prqz6Q4LEF-xu7FqGjDmh9haliI638ivcQ9XGo9_DLUrtKMh1K4UYN87j2H7vbjbJPJaw6oJ61ZF8Ngk1FSuHUvLB540QKwBJKnClMcvw-ih3CdKRdSTPRfIYlsYOkNCwEWrLwHarLTHG5HzeakMO_5usxOAvzXVDUKRIX7W62x9xKCnVl_Gp0E91dLtlU3U8j_ZZvYCj9pMvd_n_Ba82OUpFxvoXv2R2H1H8bfzztGg5E88fH6OMwoMS5OYa45nCoRFA11Bw7G8OEVqBiPzS48EY1JPjPTC5lBco0CUoeUAHPXeD7jmO-D3LhIL1-3erJ_mxP0ZeiCWUckYaFE2EliolooAHqkoVBmy5Q
- id: 0725233a85e23431681447327cbcd9d2_msg
type: message
status: completed
role: assistant
content:
- type: output_text
text: '423'
annotations: []
- type: message
role: user
content: Add one to that result. Return only the new integer.
responses:
'200':
description: 生成成功
content:
application/json:
schema:
$ref: '#/components/schemas/ResponsesResponse'
examples:
with_reasoning:
summary: 包含思考内容
value:
id: 0725233a85e23431681447327cbcd9d2
object: response
created_at: 1792405562
model: MiniMax-M3.1-Flash
status: completed
output:
- id: 0725233a85e23431681447327cbcd9d2_rs
type: reasoning
status: completed
summary: []
content:
- type: reasoning_text
text: >-
We need solve CRT. Need only n. Compute. n=3 mod7,
5 mod11,7 mod13. Product 1001. Let's solve.
Use n=3+7a. mod11: 3+7a=5 =>7a=2 mod11. inverse7=8
(56=1), a=16=5 mod11. a=5+11b. n=38+77b. mod13:
38+77b = 12+12b mod13 (77=12,38=12) equals7
=>12b=-5=8 mod13. inverse12=12, b=96 mod13 =5? 91
remainder5. n=38+385=423. Check 423 mod7 3 (420),
mod11 5 (418), mod13 7 (416). Smallest positive
423. Only n.
encrypted_content: >-
Ahlub3JtYWwtcmVhc29uaW5nLTIwMjYxMDExjehfvIf2rUinmic9ZiAutly0DD3lHqWEni9ENDziUw0nX-F6VPZLM7x1NbRWbytJNcLH9XQIa-KLYgd-it6XmXzYLvck3xOPOG9wtDdxUnjdBhuuj9dkWrkx0kbK8MFT5UreZ7EcxoRZybrs3oELf2BK_WtDUFOm_FbZXKwNbg76pTCFDoRke0Er0GMWjmi5L8GE7hGnvvBQjHXrwQ88G0L2p6STay-n1FJ5prqz6Q4LEF-xu7FqGjDmh9haliI638ivcQ9XGo9_DLUrtKMh1K4UYN87j2H7vbjbJPJaw6oJ61ZF8Ngk1FSuHUvLB540QKwBJKnClMcvw-ih3CdKRdSTPRfIYlsYOkNCwEWrLwHarLTHG5HzeakMO_5usxOAvzXVDUKRIX7W62x9xKCnVl_Gp0E91dLtlU3U8j_ZZvYCj9pMvd_n_Ba82OUpFxvoXv2R2H1H8bfzztGg5E88fH6OMwoMS5OYa45nCoRFA11Bw7G8OEVqBiPzS48EY1JPjPTC5lBco0CUoeUAHPXeD7jmO-D3LhIL1-3erJ_mxP0ZeiCWUckYaFE2EliolooAHqkoVBmy5Q
- id: 0725233a85e23431681447327cbcd9d2_msg
type: message
status: completed
role: assistant
content:
- type: output_text
text: '423'
annotations: []
output_text: '423'
usage:
input_tokens: 239
output_tokens: 190
total_tokens: 429
input_tokens_details:
cached_tokens: 174
output_tokens_details:
reasoning_tokens: 189
error: null
incomplete_details: null
instructions: null
metadata: {}
tools: null
tool_choice: auto
temperature: 1
top_p: 0.95
text:
format:
type: text
reasoning:
effort: medium
summary: null
max_output_tokens: 2048
max_tool_calls: null
parallel_tool_calls: true
previous_response_id: null
conversation: null
store: false
service_tier: null
safety_identifier: null
truncation: disabled
function_call:
summary: 触发工具调用
value:
id: 0725233e825de1d0656437967fc024e1
object: response
created_at: 1792405566
model: MiniMax-M3.1-Flash
status: completed
output:
- id: 0725233e825de1d0656437967fc024e1_msg
type: message
status: completed
role: assistant
content:
- type: output_text
text: I'll check the stock level for that product now.
annotations: []
- id: 0725233e825de1d0656437967fc024e1_fc_0
type: function_call
status: completed
call_id: call_function_izs7qcnzfksd_1
name: get_product_stock
arguments: '{"product_id": "demo-item-001"}'
output_text: I'll check the stock level for that product now.
usage:
input_tokens: 479
output_tokens: 43
total_tokens: 522
input_tokens_details:
cached_tokens: 174
error: null
incomplete_details: null
instructions: null
metadata: {}
tools:
- type: function
name: get_product_stock
description: Get the current stock level for a product.
parameters:
properties:
product_id:
description: Product identifier.
type: string
required:
- product_id
type: object
tool_choice: auto
temperature: 1
top_p: 0.95
text:
format:
type: text
reasoning:
effort: medium
summary: null
max_output_tokens: 2048
max_tool_calls: null
parallel_tool_calls: true
previous_response_id: null
conversation: null
store: false
service_tier: null
safety_identifier: null
truncation: disabled
incomplete:
summary: 达到输出上限（status=incomplete）
value:
id: 072514375f41347e5f4de3c6ceb8fd87
object: response
created_at: 1792401719
model: MiniMax-M3.1-Flash
status: incomplete
output:
- id: 072514375f41347e5f4de3c6ceb8fd87_rs
type: reasoning
status: completed
summary: []
content:
- type: reasoning_text
text: I need to explain the process step by step.
encrypted_content: >-
Ahlub3JtYWwtcmVhc29uaW5nLTIwMjYxMDExv21cbZ66-fbVryqevLX4LJWxZguAnQBWN1xlN33KB5qCPkM8IMN2pVWWIclGFlVn6ZZ0CBzkPpHqkdCVqpYbDWc3gQl2Qmf0uD4s3wo
output_text: null
usage:
input_tokens: 211
output_tokens: 8
total_tokens: 219
input_tokens_details:
cached_tokens: 195
output_tokens_details:
reasoning_tokens: 8
error: null
incomplete_details:
reason: max_output_tokens
instructions: null
metadata: {}
tools: null
tool_choice: auto
temperature: 1
top_p: 0.95
text:
format:
type: text
reasoning:
effort: max
summary: null
max_output_tokens: 8
max_tool_calls: null
parallel_tool_calls: true
previous_response_id: null
conversation: null
store: false
service_tier: null
safety_identifier: null
truncation: disabled
'400':
description: 请求参数错误
content:
application/json:
schema:
$ref: '#/components/schemas/ErrorResponse'
example:
error:
code: invalid_prompt
message: Invalid request parameters
type: invalid_request_error
'401':
description: 未认证、Token 无效或过期
content:
application/json:
schema:
$ref: '#/components/schemas/ErrorResponse'
example:
error:
code: 401
message: Invalid authentication credentials
type: authentication_error
'402':
description: 配额不足、需要充值
content:
application/json:
schema:
$ref: '#/components/schemas/ErrorResponse'
example:
error:
code: 402
message: Insufficient quota
type: insufficient_quota
'403':
description: 无权限访问该模型
content:
application/json:
schema:
$ref: '#/components/schemas/ErrorResponse'
example:
error:
code: 403
message: Model access denied
type: permission_error
'404':
description: 资源不存在
content:
application/json:
schema:
$ref: '#/components/schemas/ErrorResponse'
example:
error:
code: 404
message: Model not found
type: not_found_error
'413':
description: 请求体过大
content:
application/json:
schema:
$ref: '#/components/schemas/ErrorResponse'
example:
error:
code: 413
message: Request entity too large
type: invalid_request_error
'429':
description: 请求频率超限
content:
application/json:
schema:
$ref: '#/components/schemas/ErrorResponse'
example:
error:
code: 429
message: Rate limit exceeded
type: rate_limit_error
'500':
description: 服务器内部错误
content:
application/json:
schema:
$ref: '#/components/schemas/ErrorResponse'
example:
error:
code: 500
message: Internal server error
type: server_error
'502':
description: 网关错误
content:
application/json:
schema:
$ref: '#/components/schemas/ErrorResponse'
example:
error:
code: 502
message: Bad gateway
type: server_error
'503':
description: 服务暂时不可用
content:
application/json:
schema:
$ref: '#/components/schemas/ErrorResponse'
example:
error:
code: 503
message: Service temporarily unavailable
type: server_error
components:
schemas:
ResponsesRequest:
type: object
required:
- model
- input
properties:
model:
type: string
description: 模型 ID
enum:
- MiniMax-M3.1-Flash
example: MiniMax-M3.1-Flash
input:
type:
- string
- array
description: >-
模型输入
- **字符串形态**：整段文本按一条 `user` 消息处理
- **数组形态**：输入项列表，支持
`message`、`function_call`、`function_call_output`、`reasoning`
**多模态**：在 `message` 项的 `content` 数组里混合
`input_text`、`input_image`、`input_video` 块。
**多轮对话**：接口无状态，续接对话必须把完整历史放进数组。
items:
$ref: '#/components/schemas/InputItem'
example: Explain recursion in one sentence.
instructions:
type: string
description: 系统级指令，等价于插入在最前面的一条 system 消息，用于设定角色、语气与输出约束。
example: >-
You are a technical documentation assistant. Keep your answers
concise.
stream:
type: boolean
description: |-
是否流式返回
- `false`（默认）：一次性返回完整响应对象
- `true`：以语义化 SSE 事件推送，终止事件为 `response.completed`，**不会发送 `[DONE]`**
default: false
example: false
max_output_tokens:
type: integer
description: |-
本次生成的最大输出 token 数，**包含思考和回答**
- 推荐 **131,072**（128K），最大 **524,288**（512K）
- 取值过小时可能在思考阶段就达到上限，返回 `status: "incomplete"` 且没有 `message` 输出项
minimum: 1
maximum: 524288
example: 131072
reasoning:
type: object
description: >-
思考配置。MiniMax-M3.1-Flash 始终进行思考，思考内容以 `reasoning` 输出项返回，其 token
计入输出并按输出价计费。
properties:
effort:
type: string
description: 思考深度，默认 `max`。较高档位通常会增加思考 token 消耗和响应时间。
enum:
- low
- medium
- high
- xhigh
- max
default: max
example: high
temperature:
type: number
description: 采样温度，固定为 `1`
enum:
- 1
default: 1
example: 1
top_p:
type: number
description: 核采样阈值，默认 0.95。
minimum: 0
maximum: 1
default: 0.95
example: 0.95
tools:
type: array
description: >-
可供模型调用的工具列表，目前只支持 `function` 类型。函数由客户端执行，并把结果以
`function_call_output` 输入项回传。
items:
$ref: '#/components/schemas/Tool'
tool_choice:
type: string
description: 工具选择策略，仅支持 `auto`（由模型决定是否调用工具）
enum:
- auto
default: auto
example: auto
text:
type: object
description: 文本输出格式配置，`format.type` 仅支持 `text`
properties:
format:
type: object
properties:
type:
type: string
enum:
- text
default: text
description: 输出格式类型
metadata:
type: object
description: 请求元数据，键和值都是字符串
additionalProperties:
type: string
ResponsesResponse:
type: object
description: >-
Responses
接口的响应对象。除下列字段外，响应还会回显本次请求的配置（`tools`、`tool_choice`、`temperature`、`top_p`、`text`、`reasoning`、`max_output_tokens`
等）。
properties:
id:
type: string
description: 本次响应的唯一标识
example: 072514d63605b7a0c83f1672230c398c
object:
type: string
description: 对象类型，固定为 `response`
example: response
created_at:
type: integer
description: 创建时间，Unix 时间戳（秒）
example: 1792401878
status:
type: string
description: |-
响应状态
- `completed`：正常完成
- `incomplete`：被截断，原因见 `incomplete_details`
- `failed`：生成失败，原因见 `error`
enum:
- completed
- incomplete
- failed
example: completed
model:
type: string
description: 实际生成本次响应的模型 ID
example: MiniMax-M3.1-Flash
output:
type: array
description: >-
输出项列表，按生成顺序排列：
- `reasoning`：思考过程，id 形如 `_rs`，含 `content`（`reasoning_text`
块列表）、`summary`（空数组）、`encrypted_content`
- `message`：最终回答，id 形如 `_msg`，含 `content`（`output_text` 块列表）
- `function_call`：模型发起的函数调用，id 形如 `_fc_<序号>`，含
`call_id`、`name`、`arguments`，需客户端执行后以 `function_call_output` 回传
items:
type: object
output_text:
type:
- string
- 'null'
description: 便捷字段：所有 `output_text` 文本拼接后的结果；没有文本输出时为 `null`
error:
type:
- object
- 'null'
description: 失败原因，成功时为 `null`
incomplete_details:
type:
- object
- 'null'
description: >-
截断原因，未截断时为 `null`。`reason` 为 `max_output_tokens`（输出达到上限）或
`content_filter`（内容安全）
usage:
$ref: '#/components/schemas/Usage'
ErrorResponse:
type: object
description: 错误响应
properties:
error:
type: object
properties:
code:
type:
- integer
- string
- 'null'
description: 错误代码
example: invalid_prompt
message:
type: string
description: 错误描述
example: Invalid request parameters
type:
type: string
description: 错误类型
example: invalid_request_error
InputItem:
title: 输入项
description: 输入项：消息、思考历史、函数调用或函数执行结果。
oneOf:
- $ref: '#/components/schemas/MessageItem'
- $ref: '#/components/schemas/ReasoningItem'
- $ref: '#/components/schemas/FunctionCallItem'
- $ref: '#/components/schemas/FunctionCallOutputItem'
Tool:
type: object
title: 工具定义
required:
- type
- name
properties:
type:
type: string
description: 工具类型，目前只支持 `function`
enum:
- function
example: function
name:
type: string
description: 函数名称
example: get_product_stock
description:
type: string
description: 函数用途说明，写清楚有助于模型判断何时调用。
example: Get the current stock level for a product.
parameters:
type: object
description: 参数定义，JSON Schema 格式。
example:
type: object
properties:
product_id:
type: string
description: Product identifier.
required:
- product_id
Usage:
type: object
description: Token 使用统计信息（含缓存与思考分项）
properties:
input_tokens:
type: integer
description: 输入 token 总数（含缓存命中部分）
example: 207
input_tokens_details:
type: object
description: 输入 token 详细分项
properties:
cached_tokens:
type: integer
description: |-
命中缓存的输入 token 数
命中部分按缓存命中价计费；未命中部分 = `input_tokens - cached_tokens`，按标准输入价计费
example: 174
output_tokens:
type: integer
description: 输出 token 数量（含思考部分）
example: 32
output_tokens_details:
type: object
description: 输出 token 详细分项
properties:
reasoning_tokens:
type: integer
description: 思考产生的 token 数（已计入 `output_tokens`）
example: 9
total_tokens:
type: integer
description: 总 token 数量 = input_tokens + output_tokens
example: 239
MessageItem:
type: object
title: 消息项
description: 对话消息。回传历史助手消息时保留返回项的完整字段。
required:
- role
- content
properties:
type:
type: string
enum:
- message
default: message
id:
type: string
description: 输出项标识，回传历史时保持不变。
status:
type: string
description: 输出项状态，回传历史时保持不变。
role:
type: string
description: '消息角色，仅 `type: message` 使用。`developer` 按 `system` 处理。'
enum:
- user
- assistant
- system
- developer
example: user
content:
type:
- string
- array
description: |-
消息内容，仅 `type: message` 使用。可以是纯文本，也可以是内容块数组。
**内容块类型**：
- `input_text`：文本，字段 `text`
- `input_image`：图片，字段 `image_url`
- `input_video`：视频，字段 `video_url`
- `output_text`：assistant 历史回复中的文本
请求体最大 64 MB。
items:
$ref: '#/components/schemas/ContentPart'
example: Explain recursion in one sentence.
ReasoningItem:
type: object
title: 思考项
description: >-
上一轮返回的思考项。续答时按返回结构原样回传，包括 `id`、`status`、`summary`、`content` 和
`encrypted_content`。
required:
- type
properties:
type:
type: string
enum:
- reasoning
id:
type: string
description: 输出项标识，回传历史时保持不变。
status:
type: string
description: 输出项状态，回传历史时保持不变。
summary:
type: array
description: 思考摘要，回传历史时保持原样。
items:
type: object
content:
type: array
description: 思考内容块，回传历史时保持原样。
items:
$ref: '#/components/schemas/ReasoningContentPart'
encrypted_content:
type: string
description: 不透明的思考数据，回传历史时保持原样。
FunctionCallItem:
type: object
title: 函数调用项
description: 模型返回的函数调用，通过 `call_id` 与 `function_call_output` 项关联。
required:
- type
- call_id
- name
- arguments
properties:
type:
type: string
enum:
- function_call
id:
type: string
description: 输出项标识，回传历史时保持不变。
status:
type: string
description: 输出项状态，回传历史时保持不变。
call_id:
type: string
description: 函数调用标识，`function_call` 与 `function_call_output` 通过它配对。
name:
type: string
description: '被调用的函数名，仅 `type: function_call` 使用。'
arguments:
type: string
description: '函数调用参数，JSON 字符串，仅 `type: function_call` 使用。'
FunctionCallOutputItem:
type: object
title: 函数执行结果项
description: 客户端执行函数后的结果，通过 `call_id` 与对应的函数调用项关联。
required:
- type
- call_id
- output
properties:
type:
type: string
enum:
- function_call_output
call_id:
type: string
description: 函数调用标识，`function_call` 与 `function_call_output` 通过它配对。
output:
type: string
description: '函数执行结果，仅 `type: function_call_output` 使用。'
ContentPart:
type: object
title: 内容块
required:
- type
properties:
type:
type: string
enum:
- input_text
- input_image
- input_video
- output_text
description: 内容块类型
text:
type: string
description: 文本内容（`input_text` / `output_text`）
image_url:
description: >-
图片输入，`type: input_image` 时使用。可直接传 URL / Base64 data URL 字符串，或传对象
`{"url": "...", "detail": "low | default | high"}`
控制解析精度（越高越细致，token 越多）。支持 JPEG / PNG / GIF / WEBP，单张最大 10 MB
oneOf:
- type: string
- type: object
required:
- url
properties:
url:
type: string
description: 图片 URL 或 Base64 data URL
detail:
type: string
enum:
- low
- default
- high
default: default
description: 解析精度
video_url:
description: >-
视频输入，`type: input_video` 时使用。可直接传 URL / Base64 data URL 字符串，或传对象
`{"url": "...", "fps": 1, "detail": "default",
"max_long_side_pixel": ...}`。`fps` 为抽帧频率，默认 1，范围 [0.2, 5]，取值越低 token
越少。支持 MP4 / AVI / MOV / MKV，URL / Base64 最大 50 MB；更大的视频通过
`mm_file://{file_id}` 文件引用传入
oneOf:
- type: string
- type: object
required:
- url
properties:
url:
type: string
description: 视频 URL、Base64 data URL 或 `mm_file://{file_id}`
fps:
type: number
minimum: 0.2
maximum: 5
default: 1
description: 抽帧频率
detail:
type: string
enum:
- low
- default
- high
default: default
description: 解析精度
max_long_side_pixel:
type: integer
minimum: 1
description: 视频帧最长边像素限制
ReasoningContentPart:
type: object
title: 思考内容块
required:
- type
- text
properties:
type:
type: string
enum:
- reasoning_text
text:
type: string
description: 思考文本。
securitySchemes:
bearerAuth:
type: http
scheme: bearer
description: |-
所有接口均需要使用 Bearer Token 进行认证
**获取 API Key**：
访问 [API Key 管理页面](https://evolink.ai/dashboard/keys) 获取您的 API Key
**使用时在请求头中添加**：
```
Authorization: Bearer YOUR_API_KEY
```
````
This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.