Back to Models

Qwen3 Coder 480B

qwenqwen/qwen3-coder-480b

API guide

Chat completion

Use the standard OpenAI-compatible chat completions endpoint. This model is strongest when the prompt contains concrete files, failing tests, or explicit coding constraints.

cURL
curl https://api.therouter.ai/v1/chat/completions \
  -H "Authorization: Bearer $THEROUTER_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "qwen/qwen3-coder-480b",
    "messages": [
      {"role": "system", "content": "You are a senior software engineer. Return concise patches and explain tradeoffs."},
      {"role": "user", "content": "Refactor this TypeScript function to remove duplication and preserve behavior: ..."}
    ],
    "temperature": 0.7,
    "max_tokens": 4096
  }'

Streaming

Stream long code explanations and patch plans so IDE users see progress before the full answer finishes.

cURL
curl https://api.therouter.ai/v1/chat/completions \
  -H "Authorization: Bearer $THEROUTER_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "qwen/qwen3-coder-480b",
    "messages": [{"role": "user", "content": "Explain the failure mode and propose a minimal patch."}],
    "stream": true
  }'

Tool calling

Qwen3-Coder is tuned for agentic workflows. Use tools for repository search, test execution, issue lookup, and patch application rather than asking the model to guess state.

cURL
curl https://api.therouter.ai/v1/chat/completions \
  -H "Authorization: Bearer $THEROUTER_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "qwen/qwen3-coder-480b",
    "messages": [{"role": "user", "content": "Find the file that defines billing reconciliation and summarize the invariant."}],
    "tools": [{
      "type": "function",
      "function": {
        "name": "repo_search",
        "description": "Search repository files by query",
        "parameters": {
          "type": "object",
          "properties": {"query": {"type": "string"}},
          "required": ["query"]
        }
      }
    }],
    "tool_choice": "auto"
  }'

JSON mode

Use JSON mode for code review automation, structured issue triage, and migration inventories.

cURL
curl https://api.therouter.ai/v1/chat/completions \
  -H "Authorization: Bearer $THEROUTER_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "qwen/qwen3-coder-480b",
    "messages": [
      {"role": "system", "content": "Return only JSON with keys risk, files, tests."},
      {"role": "user", "content": "Review this diff for production risk: ..."}
    ],
    "response_format": {"type": "json_object"}
  }'
Fact ledger β€” every claim on this page traces here
sourceURLretrieved
Architecturehuggingface.co β†—2026-06-02verified
Parametershuggingface.co β†—2026-06-02verified
Expertshuggingface.co β†—2026-06-02verified
Native contextqwenlm.github.io β†—2026-06-02verified
Output modehuggingface.co β†—2026-06-02verified
Training emphasisqwenlm.github.io β†—2026-06-02verified
SWE-bench Verifiedtogether.ai β†—2026-06-02to verify
Aider Polyglotmedium.com β†—2026-06-02single source
Agentic Coding / Browser-Use / Tool-Useqwenlm.github.io β†—2026-06-02verified
Qwen3-Coder-Next technical report compares the new 80B/3B model to Qwen3-Coder-480Barxiv.org β†—2026-06-02verified
Qwen launches Qwen3-Coder-480B-A35B-Instruct and Qwen Codeqwenlm.github.io β†—2026-06-02verified
Hugging Face card publishes architecture and non-thinking-mode detailshuggingface.co/Qwen β†—2026-06-02verified
Is Qwen3-Coder-480B a thinking model?huggingface.co β†—2026-06-02to verify
How much context does it support?qwenlm.github.io β†—2026-06-02to verify
Can I use it from Cursor, Cline, or Claude Code-style tools?qwenlm.github.io β†—2026-06-02to verify
What makes it different from qwen/qwen3-coder-30b?github.com β†—2026-06-02to verify
Can I self-host Qwen3-Coder-480B?huggingface.co β†—2026-06-02to verify
Help & contact