> ## Documentation Index > Fetch the complete documentation index at: https://platform.minimax.io/docs/llms.txt > Use this file to discover all available pages before exploring further. # Text Generation > Use this API to create chat completions. ## OpenAPI ````yaml /api-reference/text/api/openapi.json POST /v1/text/chatcompletion_v2 openapi: 3.1.0 info: title: MiniMax Text API description: >- MiniMax text generation API with support for chat completion and streaming output license: name: MIT version: 1.0.0 servers: - url: https://api.minimax.io security: - bearerAuth: [] paths: /v1/text/chatcompletion_v2: post: tags: - Text Generation summary: Text Generation V2 operationId: chatCompletionV2 parameters: - name: Content-Type in: header required: true description: >- The media type of the request body. Must be set to `application/json` to ensure the data is sent in JSON format. schema: type: string enum: - application/json default: application/json requestBody: content: application/json: schema: $ref: '#/components/schemas/ChatCompletionReq' examples: Request: value: model: MiniMax-M3 messages: - role: system name: MiniMax AI - role: user name: user content: hello Stream: value: model: MiniMax-M3 messages: - role: system name: MiniMax AI - role: user name: user content: hello stream: true Image: value: model: MiniMax-Text-01 messages: - role: system name: MiniMax AI content: >- MM Smart Assistant is a large language model independently developed by MiniMax, without relying on other products' APIs. MiniMax is a Chinese technology company dedicated to research in large-scale models. - role: user name: User content: - type: text text: What does this picture represent? - type: image_url image_url: url: >- https://cdn.hailuoai.com/prod/2024-09-18-16/user/multi_chat_file/9c0b5c14-ee88-4a5b-b503-4f626f018639.jpeg required: true responses: '200': description: '' content: application/json: schema: $ref: '#/components/schemas/ChatCompletionResp' examples: Request: value: id: 04ecb5d9b1921ae0fb0e8da9017a5474 choices: - finish_reason: stop index: 0 message: content: Hello! How can I assist you? role: assistant name: MiniMax AI audio_content: '' reasoning_content: ...omitted created: 1755153113 model: MiniMax-M3 object: chat.completion usage: total_tokens: 249 total_characters: 0 prompt_tokens: 26 completion_tokens: 223 completion_tokens_details: reasoning_tokens: 214 input_sensitive: false output_sensitive: false input_sensitive_type: 0 output_sensitive_type: 0 output_sensitive_int: 0 base_resp: status_code: 0 status_msg: '' Stream: value: - id: 02ff7eb7fe6fb505b9d5cb6945a1a98b choices: - index: 0 delta: content: Okay role: assistant created: 1722829751 model: MiniMax-M3 object: chat.completion.chunk output_sensitive: false input_sensitive_type: 0 output_sensitive_type: 0 - id: 02ff7eb7fe6fb505b9d5cb6945a1a98b choices: - finish_reason: stop index: 0 delta: content: Hello! How can I assist you today? role: assistant created: 1722829751 model: MiniMax-M3 object: chat.completion.chunk output_sensitive: false input_sensitive_type: 0 output_sensitive_type: 0 - id: 02ff7eb7fe6fb505b9d5cb6945a1a98b choices: - finish_reason: stop index: 0 message: content: Hello! How can I assist you today? role: assistant created: 1722829751 model: MiniMax-M3 object: chat.completion usage: total_tokens: 73 input_sensitive: false output_sensitive: false input_sensitive_type: 0 output_sensitive_type: 0 base_resp: status_code: 0 status_msg: '' Image: value: content: >- This picture represents a cute and whimsical scene where a hamster is depicted as if it is having a cup of tea. The hamster is placed on a small wooden table with a miniature teacup and saucer, creating a charming and playful atmosphere. The setting includes small potted plants and a clock in the background, adding to the cozy and homely feel of the image. role: assistant name: MiniMax AI audio_content: '' text/event-stream: schema: $ref: '#/components/schemas/ChatCompletionResp' examples: Stream: value: - id: 02ff7eb7fe6fb505b9d5cb6945a1a98b choices: - index: 0 delta: content: Okay role: assistant created: 1722829751 model: MiniMax-M3 object: chat.completion.chunk output_sensitive: false input_sensitive_type: 0 output_sensitive_type: 0 - id: 02ff7eb7fe6fb505b9d5cb6945a1a98b choices: - finish_reason: stop index: 0 delta: content: Hello! How can I assist you today? role: assistant created: 1722829751 model: MiniMax-M3 object: chat.completion.chunk output_sensitive: false input_sensitive_type: 0 output_sensitive_type: 0 - id: 02ff7eb7fe6fb505b9d5cb6945a1a98b choices: - finish_reason: stop index: 0 message: content: Hello! How can I assist you today? role: assistant created: 1722829751 model: MiniMax-M3 object: chat.completion usage: total_tokens: 73 input_sensitive: false output_sensitive: false input_sensitive_type: 0 output_sensitive_type: 0 base_resp: status_code: 0 status_msg: '' components: schemas: ChatCompletionReq: type: object required: - model - messages properties: model: type: string description: >- Model ID. Options: `MiniMax-M3`, `MiniMax-M2.7`, `MiniMax-M2.7-highspeed`, `MiniMax-M2.5`, `MiniMax-M2.5-highspeed`, `MiniMax-M2.1`, `MiniMax-M2`. `MiniMax-M3`, `MiniMax-M2.7`, `MiniMax-M2.7-highspeed`, `MiniMax-M2.5`, `MiniMax-M2.5-highspeed`, `MiniMax-M2.1` and `MiniMax-M2` are reasoning models, recommended with streaming output for best performance. enum: - MiniMax-M3 - MiniMax-M2.7 - MiniMax-M2.7-highspeed - MiniMax-M2.5 - MiniMax-M2.5-highspeed - MiniMax-M2.1 - MiniMax-M2 stream: type: boolean description: >- Whether to enable streaming. Defaults to `false`. If `true`, the response will be returned in multiple chunks. default: false max_tokens: type: integer format: int64 description: >- Deprecated in favor of `max_completion_tokens`. Maximum number of tokens allowed in the generated output. Content beyond this limit will be truncated. If the generation is cut off due to length limits, consider increasing this value. Default: - `MiniMax-M2`: 10240 - `MiniMax-M1`: 8192 - `MiniMax-Text-01`: 2048 minimum: 1 deprecated: true max_completion_tokens: type: integer format: int64 description: >- Maximum number of tokens allowed in the generated output. Content beyond this limit will be truncated. If the generation is cut off due to length limits, consider increasing this value. Default: - `MiniMax-M2`: 10240 - `MiniMax-M1`: 8192 - `MiniMax-Text-01`: 2048 minimum: 1 temperature: type: number format: double description: >- Controls randomness of the output. Range: (0, 1]. Higher values will make the output more random, while lower values will make it more focused and deterministic. Default values: - `MiniMax-M2`: recommended 1.0 -`MiniMax-M1`: 1.0 (recommended range [0.8, 1.0]) - `MiniMax-Text-01`: 0.1 (can be raised to 0.7–1.0 for creative tasks). minimum: 0 exclusiveMinimum: 0 maximum: 1 top_p: type: number format: double description: |- Sampling strategy that affects output diversity. Range: (0, 1] - `MiniMax-M2` recommended 0.95 minimum: 0 exclusiveMinimum: 0 maximum: 1 default: 0.95 messages: type: array description: List of messages containing conversation history items: $ref: '#/components/schemas/Message' tool_choice: type: string description: |- Controls how the model uses tools: - `none`: never call tools - `auto`: decide automatically (default) enum: - none - auto default: auto tools: type: array description: >- List of tools available for the model to call. For more details, see [M2 Tool Use & Interleaved Thinking](/guides/text-m3-function-call) items: type: object required: - type - function properties: type: type: string description: Tool type. Currently only `function` is supported. enum: - function function: type: object required: - name - description - parameters description: Function definition. properties: name: type: string description: Function name. description: type: string description: >- Description of the function's purpose, behavior, and parameters. parameters: type: object description: Function parameters. response_format: type: object description: |- Specifies the output format (only supported by `MiniMax-Text-01`). Example: `{ "type": "json_schema", "json_schema": {...} }` required: - type - json_schema properties: type: type: string description: Format type. Only `json_schema` is supported. enum: - json_schema json_schema: type: object description: Schema definition for structured output. required: - name - schema properties: name: type: string description: 'Schema name. Max length: 64. Must match regex: `\w`.' maxLength: 64 pattern: ^\w+$ description: type: string description: >- Description of the schema, helps model understand output format. schema: type: object description: JSON Schema definition. required: - type - properties properties: type: type: string description: Must be `object`. enum: - object properties: type: object description: >- Define each field in the output object. Supports types: String, Array, Enum, Number, Integer, Object, Boolean. **Example:** ```json "json_schema": { "name": "user_analysis", "description": "User behavior analysis result", "schema": { "type": "object", "properties": { "analysis": { "type": "string", "description": "Content analysis based on rules" }, "summary_words": { "type": "array", "items": { "type": "string" }, "description": "Content keywords summary, limited to three words" }, "content_result": { "type": "string", "enum": [ "Pornography/Obscenity Violation", "Advertising/Spam Violation", "Illegal Content Violation", "Political Content Violation", "Self-harm/Suicide Violation", "Underage Violation", "Abuse/Insult Violation", "Other Violations", "Normal" ], "description": "Content classification result" }, "risk_words": { "type": "array", "items": { "type": "string" }, "description": "Keywords indicating violations, emojis converted to text" }, "correct_rate": { "type": "number", "description": "Credibility score, from 0.00 to 1.00" }, "content_risk_level": { "type": "string", "enum": ["high", "medium", "low"], "description": "Risk concentration level in content" } }, "required": [ "analysis", "summary_words", "content_result", "risk_words", "correct_rate", "content_risk_level" ] } } ``` stream_options: type: object description: Options for streaming mode. properties: include_usage: type: boolean description: >- Defaults to `false`. If `true`, the final chunk will include complete `usage` statistics. default: false mask_sensitive_info: type: boolean description: >- Mask (replace with ***) content in the output that involves private information, including but not limited to email, domain, link, ID number, home address, etc. Defaults `false`. default: false example: model: MiniMax-M3 messages: - role: system name: MiniMax AI - role: user name: user content: hello ChatCompletionResp: type: object properties: id: type: string description: Unique ID of this response. choices: type: array description: List of response choices. items: type: object properties: finish_reason: type: string description: |- Reason the generation stopped: - `stop`: completed naturally - `length`: reached max_tokens limit - `tool_calls`: tool call triggered enum: - stop - length - tool_calls index: type: integer description: Index of the choice (starting at 0). message: type: object description: The complete reply generated by the model. required: - content - role properties: content: type: string description: Text reply content. role: type: string description: 'Role of the message. Fixed value: `assistant`.' enum: - assistant tool_calls: type: array description: List of tools requested by the model. items: type: object required: - id - type - function properties: id: type: string description: Unique tool call ID. type: type: string description: Always `function`. enum: - function function: type: object description: Function calling information. required: - name - arguments properties: name: type: string description: name of function. arguments: type: string description: JSON string of arguments. created: type: integer format: int64 description: Unix timestamp (seconds) when the response was created. model: type: string description: Model ID used for this request. object: type: string description: |- Object type: - `chat.completion`: non-streaming - `chat.completion.chunk`: streaming enum: - chat.completion - chat.completion.chunk usage: $ref: '#/components/schemas/Usage' input_sensitive: type: boolean description: >- Whether the input contained sensitive content. See `input_sensitive_type` for details on the type of sensitive content. input_sensitive_type: type: integer format: int64 description: |- Type of sensitive content in input. Possible values: `1`: severe violation `2`: pornography `3`: advertising `4`: prohibited `5`: abusive `6`: terrorism/violence `7`: other output_sensitive: type: boolean description: Whether the output contained sensitive content. output_sensitive_type: type: integer format: int64 description: Type of sensitive content in output. base_resp: type: object description: Error status codes and details. properties: status_code: type: integer format: int64 description: >- Error code. Examples: `1000`: Unknown error `1001`: Request timeout `1002`: Rate limited `1004`: Authentication failed `1008`: Insufficient balance `1013`: Internal server error `1027`: Invalid output content `1039`: Token limit exceeded `2013`: Invalid parameters For more information, please refer to the [Error Code Reference](/api-reference/errorcode). status_msg: type: string description: Error details. example: id: 04ecb5d9b1921ae0fb0e8da9017a5474 choices: - finish_reason: stop index: 0 message: content: Hello! How can I assist you? role: assistant name: MiniMax AI audio_content: '' reasoning_content: ...omitted created: 1755153113 model: MiniMax-M3 object: chat.completion usage: total_tokens: 249 total_characters: 0 prompt_tokens: 26 completion_tokens: 223 completion_tokens_details: reasoning_tokens: 214 input_sensitive: false output_sensitive: false input_sensitive_type: 0 output_sensitive_type: 0 output_sensitive_int: 0 base_resp: status_code: 0 status_msg: '' Message: type: object required: - role - content properties: role: type: string enum: - system - user - assistant - tool description: |- The role of the message sender: - `system`: set the model's role and behavior - `user`: user input - `assistant`: model responses, may include tool calls - `tool`: returned result from a tool call name: type: string description: >- Name of the sender. Optional parameter. Required if multiple senders share the same role. content: description: 'Message content. Supports two types: `string` and `object[]`' oneOf: - type: string description: Plain `string` message content - type: array description: Multimodal message content (`array`) items: type: object required: - type properties: type: type: string enum: - text - image_url description: 'Content type. Options: `text`, `image_url`.' text: type: string description: Text content. image_url: type: object required: - url properties: url: type: string description: Public image URL or Base64-encoded Data. description: Image input. tool_calls: type: array description: >- List of tools the model decides to invoke (appears when `role=assistant`). items: type: object required: - id - type - function properties: id: type: string description: Unique ID of the tool call. type: type: string description: 'Tool type. Fixed value: `function`.' enum: - function function: type: object required: - name - arguments description: Details of the function call. properties: name: type: string description: Name of the function to be called. arguments: type: string description: JSON string containing the function call arguments. Usage: type: object description: Token usage statistics for this request. properties: total_tokens: type: integer description: Total number of tokens consumed. securitySchemes: bearerAuth: type: http scheme: bearer bearerFormat: JWT description: >- `HTTP: Bearer Auth` - Security Scheme Type: http - HTTP Authorization Scheme: `Bearer API_key`, can be found in [Account Management>API Keys](https://platform.minimax.io/user-center/basic-information/interface-key). ````