---
title: "qwen3-30b-a3b-fp8: Chat completions"
method: POST
path: "/chat/completions"
tags: ["qwen3-30b-a3b-fp8"]
---

# qwen3-30b-a3b-fp8: Chat completions

`POST /chat/completions`

## Headers

- `x-project-id` string

## Request body

- object
  - `model` 'qwen3-30b-a3b-fp8', required — Model identifier (fixed for this playground). Use request examples to change use cases.
  - `messages` ChatMessage[], required
    - `role` 'system' | 'user' | 'assistant' | 'tool' | 'developer', required — A `developer` role is normalized to `system`; unrecognized roles are treated as `user`.
    - `content` string, required
  - `max_tokens` integer — Maximum number of tokens to generate in the response.
  - `metadata` object

## Response `200`

Success

- ChatCompletionResponse — OpenAI Chat Completions response. For classification models, `choices[0].message.content` is a JSON-stringified classification result.
  - `id` string
  - `object` 'chat.completion'
  - `created` integer
  - `model` string
  - `choices` object[]
    - `index` integer
    - `finish_reason` string
    - `message` object
      - `role` 'assistant'
      - `content` string
      - `annotations` unknown[]
        - unknown
  - `usage` ChatUsage
    - `prompt_tokens` integer
    - `completion_tokens` integer
    - `total_tokens` integer
    - `prompt_tokens_details` object
    - `completion_tokens_details` object

## Other responses

- `400` — Bad request
- `401` — Unauthorized
- `403` — Forbidden
- `420` — Insufficient quota
- `500` — Internal server error

---

[API](https://skmtc.dev/zerogpu/apis/zerogpu-batch-files-api.md) · [All operations](https://skmtc.dev/zerogpu/apis/zerogpu-batch-files-api/llms.txt) · [OpenAPI document](https://skmtc-service-production.skmtc.workers.dev/v1/apis/zerogpu/zerogpu-batch-files-api/revisions/f7371035d1a2/schema)