Qwen: Qwen3 VL 235B A22B Thinking
Analysis Summary
Qwen3 VL 235B A22B Thinking is a large mixture-of-experts vision-language model from Alibaba's Qwen team, combining extended thinking with multimodal input. Math benchmarks are strong at 88.3, and GPQA at 0.772 shows reasonable scientific reasoning. Tool use, function calling, and vision are all supported, giving it reasonable breadth for a thinking-mode model.
However, the agentic index at 32.8 and coding index at 20.9 are below what businesses need for autonomous agents or serious software engineering. The thinking mode also raises output costs to $2.60 per million tokens, which limits its appeal for high-volume use. It suits analytical and research-adjacent tasks where vision and extended reasoning matter more than coding or agentic reliability.
A -4 regional accessibility adjustment applies. For businesses already integrated with Qwen's API, it is a capable multimodal reasoning tool; for others, major providers offer more accessible alternatives at comparable or better performance.
Assessed June 6, 2026
Editorial notes
Qwen3 VL 235B A22B Thinking from Qwen delivers strong math and vision reasoning with tool use and function calling, though agentic scores are limited and the -4 regional penalty applies.
Rankings consider pricing, capabilities, benchmarks, and real-world applicability and are refreshed as new models launch. Feedback?
Performance Profile
How Qwen: Qwen3 VL 235B A22B Thinking compares
Qwen: Qwen3 VL 235B A22B Thinking ranks #123 of 378 AI models we track for overall intelligence, #140 of 315 for coding, #132 of 289 for agentic tasks. Its 131K-token context window is larger than 60% of the models we list. At $0.26 per million input tokens it is cheaper than 45% of comparable models.
About Qwen: Qwen3 VL 235B A22B Thinking
Qwen3-VL-235B-A22B Thinking is a multimodal model that unifies strong text generation with visual understanding across images and video. The Thinking model is optimized for multimodal reasoning in STEM and math..
Capabilities
Performance Indices
Source: Artificial Analysis
Benchmark Scores
Intelligence
Technical
Content
Benchmark data from Artificial Analysis and Hugging Face
How does Qwen: Qwen3 VL 235B A22B Thinking stack up?
Compare side-by-side with other professional models.
Model Information
| OpenRouter ID |
qwen/qwen3-vl-235b-a22b-thinking
|
| Provider | qwen |
| Release Date | September 23, 2025 |
| Context Length | 131,072 tokens |
| Max Completion | 32,768 tokens |
| Status | Active |
Pricing
| Token Type | Cost per 1M tokens | Cost per 1K tokens |
|---|---|---|
| Input | $0.26 | $0.000260 |
| Output | $2.60 | $0.002600 |
Live Performance
Live endpoint metrics, refreshed every 30 minutes.
Leaderboard Categories
External Resources
Frequently asked questions about Qwen: Qwen3 VL 235B A22B Thinking
How much does Qwen: Qwen3 VL 235B A22B Thinking cost?
Qwen: Qwen3 VL 235B A22B Thinking costs $0.26 per million input tokens and $2.60 per million output tokens.
What is the context window of Qwen: Qwen3 VL 235B A22B Thinking?
Qwen: Qwen3 VL 235B A22B Thinking has a context window of 131,072 tokens (131K).
Is Qwen: Qwen3 VL 235B A22B Thinking good for coding?
On our coding benchmark index, Qwen: Qwen3 VL 235B A22B Thinking ranks #140 of 315 models, placing it in the broader range of the field for code generation and debugging.
What can Qwen: Qwen3 VL 235B A22B Thinking do?
Qwen: Qwen3 VL 235B A22B Thinking supports image/vision input, tool use, and function calling.
Who created Qwen: Qwen3 VL 235B A22B Thinking?
Qwen: Qwen3 VL 235B A22B Thinking is developed by Qwen and was released on September 23, 2025.
