Back to Models
Qwen3 30B Thinking
qwenqwen/qwen3-30b-thinking
How TheRouter serves this differently from the vendor
As the vendor operates it
Alibaba publishes Qwen3-30B-A3B as open model artifacts and documentation; host deployments may expose tokenizer templates, thinking controls, or snapshot labels according to their serving stack.
On TheRouter
TheRouter serves qwen/qwen3-30b-thinking as a text-only, OpenAI-compatible chat-completions route with the supported parameters declared in standard-models.yaml.
API guide
Chat completion
cURL
curl https://api.therouter.ai/v1/chat/completions \
-H "Authorization: Bearer $THEROUTER_API_KEY" \
-H "Content-Type: application/json" \
-d '{"model":"qwen/qwen3-30b-thinking","messages":[{"role":"user","content":"Compare these two incident summaries and identify the likely root cause."}]}'Reasoning prompt
Use a bounded final-answer instruction so the model spends reasoning effort on the comparison but returns concise production output.
cURL
curl https://api.therouter.ai/v1/chat/completions \
-H "Authorization: Bearer $THEROUTER_API_KEY" \
-H "Content-Type: application/json" \
-d '{"model":"qwen/qwen3-30b-thinking","temperature":0.2,"messages":[{"role":"user","content":"Reason through the tradeoffs, then return only: recommendation, risks, and first test."}]}'Fact ledger — every claim on this page traces here
| source | URL | retrieved | |
|---|---|---|---|
| Architecture | huggingface.co/Qwen ↗ | 2026-08-06 | verified |
| Context window on TheRouter | therouter.ai model configuration ↗ | 2026-08-06 | verified |
| Maximum completion | therouter.ai model configuration ↗ | 2026-08-06 | verified |
| Modalities | therouter.ai model configuration ↗ | 2026-08-06 | verified |
| Supported request parameters | therouter.ai model configuration ↗ | 2026-08-06 | verified |
| License | huggingface.co/Qwen ↗ | 2026-08-06 | verified |
| Training cutoff | — | — | unknown |
| Qwen3 family benchmark set | qwenlm.github.io ↗ | 2026-08-06 | unknown |
| AIME | qwenlm.github.io ↗ | 2026-08-06 | unknown |
| LiveCodeBench | qwenlm.github.io ↗ | 2026-08-06 | unknown |
| Qwen3 30B Thinking is available as a TheRouter reasoning route | therouter.ai model configuration ↗ | 2026-08-06 | verified |
| Is qwen/qwen3-30b-thinking the same as qwen/qwen3-30b-a3b-thinking-2507? | therouter.ai model configuration ↗ | 2026-08-06 | to verify |
| Does this TheRouter route support tool calling or JSON schema output? | therouter.ai model configuration ↗ | 2026-08-06 | to verify |
| Why is the snippet score not marked verified? | therouter.ai model configuration ↗ | 2026-08-06 | to verify |