> ## Documentation Index > Fetch the complete documentation index at: https://docs.modellix.ai/llms.txt > Use this file to discover all available pages before exploring further. # Chat Completions > OpenAI Chat Completions–compatible endpoint on the Modellix LLM gateway. Accepts a synchronous request with `model` (provider/name format) and an OpenAI-style `messages` array; returns a chat.completion JSON object by default, or SSE (`text/event-stream`) when `stream=true`. Use this path for the OpenAI SDK, Codex, and most OpenAI-compatible chat clients. Do not send Responses fields (`input`, `max_output_tokens`) or Anthropic Messages-only shapes on this endpoint—use `/responses` or `/messages` instead. Optional fields follow [OpenAI Chat Completions](https://platform.openai.com/docs/api-reference/chat); support depends on the selected model. ## OpenAPI ````yaml /llm/llm.json post /chat/completions openapi: 3.1.0 info: title: Modellix LLM API description: >- Public API of the Modellix LLM Gateway (llm-bff): OpenAI-compatible Chat Completions and Responses, plus Anthropic-compatible Messages. Optional fields follow the corresponding official protocols; this document lists commonly used core fields only. See `docs/api/llm.md` for the human-readable guide. version: '2026-08-04' contact: name: Modellix servers: - url: https://llm.modellix.ai/v1 description: Production security: - BearerAuth: [] - ApiKeyAuth: [] tags: - name: Chat Completions description: OpenAI-compatible Chat Completions - name: Responses description: OpenAI-compatible Responses - name: Messages description: Anthropic-compatible Messages paths: /chat/completions: post: tags: - Chat Completions summary: Chat Completions description: >- OpenAI Chat Completions–compatible endpoint on the Modellix LLM gateway. Accepts a synchronous request with `model` (provider/name format) and an OpenAI-style `messages` array; returns a chat.completion JSON object by default, or SSE (`text/event-stream`) when `stream=true`. Use this path for the OpenAI SDK, Codex, and most OpenAI-compatible chat clients. Do not send Responses fields (`input`, `max_output_tokens`) or Anthropic Messages-only shapes on this endpoint—use `/responses` or `/messages` instead. Optional fields follow [OpenAI Chat Completions](https://platform.openai.com/docs/api-reference/chat); support depends on the selected model. operationId: createChatCompletion parameters: - $ref: '#/components/parameters/SessionIdHeader' requestBody: required: true description: >- OpenAI Chat Completions request body. Requires model and messages; optional stream, max_tokens, max_completion_tokens, temperature, and n. content: application/json: schema: $ref: '#/components/schemas/ChatCompletionRequest' examples: nonStream: summary: Non-streaming value: model: openai/gpt-5.6-sol stream: false max_tokens: 256 messages: - role: user content: Introduce yourself in one sentence stream: summary: Streaming value: model: openai/gpt-5.6-sol stream: true max_tokens: 256 messages: - role: user content: ping responses: '200': description: Success. Non-streaming returns JSON; streaming returns SSE. content: application/json: schema: $ref: '#/components/schemas/ChatCompletionResponse' text/event-stream: schema: type: string description: >- SSE: `data: {chat.completion.chunk}` lines, terminated by `data: [DONE]` '400': $ref: '#/components/responses/BadRequest' '401': $ref: '#/components/responses/Unauthorized' '402': $ref: '#/components/responses/PaymentRequired' '404': $ref: '#/components/responses/NotFound' '429': $ref: '#/components/responses/RateLimited' 5XX: $ref: '#/components/responses/UpstreamError' components: parameters: SessionIdHeader: name: X-Mdlx-Session-Id in: header required: false description: >- Optional session affinity header. Length 8–128; alphanumeric, `-`, and `_` only. schema: type: string minLength: 8 maxLength: 128 pattern: ^[A-Za-z0-9_-]+$ example: my-conversation-001 schemas: ChatCompletionRequest: type: object required: - model - messages additionalProperties: true description: Core fields below; other fields follow OpenAI Chat Completions. properties: model: $ref: '#/components/schemas/ModelId' messages: type: array minItems: 1 description: >- OpenAI Chat Completions message list (min 1). Each item has role and content; roles include system, user, assistant, and tool. items: $ref: '#/components/schemas/ChatMessage' stream: type: boolean default: false description: >- Default false. When true, returns SSE chat.completion.chunk events ending with data: [DONE]. max_tokens: type: integer minimum: 1 description: Optional maximum number of tokens to generate. Model-dependent. max_completion_tokens: type: integer minimum: 1 description: Optional generation cap preferred by some newer OpenAI-style models. temperature: type: number description: Sampling temperature for Chat Completions. Model-dependent. 'n': type: integer minimum: 1 description: Number of chat completion choices to generate. ChatCompletionResponse: type: object additionalProperties: true description: Shape follows OpenAI chat.completion properties: id: type: string description: Unique identifier for this chat completion. object: type: string example: chat.completion description: Object type. Typically chat.completion for non-streaming responses. created: type: integer description: Unix timestamp (seconds) when the chat completion was created. model: type: string description: Model ID that produced this chat completion. choices: type: array description: List of chat completion choices. items: type: object additionalProperties: true description: >- A single OpenAI chat completion choice (message, finish_reason, index, and related fields). usage: $ref: '#/components/schemas/Usage' ModelId: type: string description: Model ID in `provider/name` form examples: - openai/gpt-5.5 - openai/gpt-5.6-sol - anthropic/claude-sonnet-5 - google/gemini-3.6-flash ChatMessage: type: object required: - role - content additionalProperties: true description: A single OpenAI Chat Completions message with role and content. properties: role: type: string description: e.g. system / user / assistant / tool content: description: Text string or content-part array oneOf: - type: string - type: array items: type: object additionalProperties: true description: OpenAI Chat Completions content part object Usage: type: object additionalProperties: true description: OpenAI chat.completion token usage counts. properties: prompt_tokens: type: integer description: Number of tokens in the prompt for Chat Completions. completion_tokens: type: integer description: Number of tokens in the generated completion. total_tokens: type: integer description: Total tokens used (prompt plus completion). ErrorBody: type: object required: - error description: >- Gateway error envelope shared by Chat Completions, Responses, and Messages. properties: error: type: object required: - message - type additionalProperties: true description: Error details object. properties: message: type: string description: Human-readable error message. type: type: string description: >- Error type string (for example invalid_request_error, rate_limit_exceeded, insufficient_quota). examples: - invalid_request_error - api_error - rate_limit_exceeded - request_limited - insufficient_quota code: type: - string - 'null' description: Optional machine-readable error code, or null. param: type: - string - 'null' description: Optional parameter name related to the error, or null. responses: BadRequest: description: Invalid request parameters content: application/json: schema: $ref: '#/components/schemas/ErrorBody' Unauthorized: description: >- Missing or invalid API key, or conflicting Authorization and x-api-key values content: application/json: schema: $ref: '#/components/schemas/ErrorBody' example: error: message: invalid API key type: invalid_request_error PaymentRequired: description: Insufficient account balance content: application/json: schema: $ref: '#/components/schemas/ErrorBody' example: error: message: insufficient balance, please recharge your account type: insufficient_quota NotFound: description: Unknown path, or the selected model is unavailable content: application/json: schema: $ref: '#/components/schemas/ErrorBody' examples: modelUnavailable: summary: Model unavailable value: error: message: the selected model is unavailable type: invalid_request_error RateLimited: description: >- Rate limited (RPM or other request limits), or the selected model is temporarily unavailable (e.g. upstream deployment cooldown) headers: X-RateLimit-Limit: description: Maximum number of requests allowed in the current rate-limit window. schema: type: integer X-RateLimit-Remaining: description: Number of requests remaining in the current rate-limit window. schema: type: integer X-RateLimit-Reset: description: Unix timestamp (seconds) when the current rate-limit window resets. schema: type: integer Retry-After: schema: type: integer description: >- May be returned when the model is temporarily unavailable; value in seconds content: application/json: schema: $ref: '#/components/schemas/ErrorBody' examples: rpm: summary: RPM exceeded value: error: message: rate limit exceeded type: rate_limit_exceeded modelCooldown: summary: Model temporarily unavailable value: error: message: >- the selected model is temporarily unavailable, please retry later or try another model type: api_error UpstreamError: description: >- Temporary service unavailability or upstream error (client-facing message may be sanitized) content: application/json: schema: $ref: '#/components/schemas/ErrorBody' securitySchemes: BearerAuth: type: http scheme: bearer bearerFormat: API Key description: >- Authorization: Bearer (OpenAI SDK, Codex, OpenCode, etc.) ApiKeyAuth: type: apiKey in: header name: x-api-key description: >- x-api-key: (Anthropic SDK, Claude Code, etc.). Equivalent to Bearer; if both are present they must be the same key. ````