# Qwen3 235B A22B Thinking 2507 API prices

Qwen3 235B A22B Thinking 2507 (qwen) — 262k context, 131k max output.
7 metered per-token offers from 9 provider(s), in USD per 1M tokens.

| Provider | Input $/1M | Output $/1M | Cache read $/1M | Context |
| --- | --- | --- | --- | --- |
| submodel | $0.2 | $0.6 | — | 262k |
| OpenRouter | $0.23 | $2.3 | — | 131k |
| Hugging Face | $0.3 | $3 | — | 262k |
| Jiekou.AI | $0.3 | $3 | — | 131k |
| LLM Gateway | $0.3 | $3 | — | 131k |
| NovitaAI | $0.3 | $3 | — | 131k |
| Vercel AI Gateway | $0.4 | $4 | — | 131k |

## Summary

- Cheapest input: $0.2 per 1M tokens (submodel)
- Cheapest output: $0.6 per 1M tokens (submodel)
- First-party: not listed separately
- Free offers: ModelScope
- Subscription plans (not per-token): none
- Released: 2025-07-25
- Inputs: image, pdf, text

HTML page: https://llmprice.gitlab.io/models/qwen3-235b-a22b-thinking-2507/
Full dataset: https://llmprice.gitlab.io/data/catalog.json