qwen
Qwen: Qwen3 VL 30B A3B Thinking
qwen/qwen3-vl-30b-a3b-thinking
Our take
AI-assisted analysis based on the sources below. Suitability is an editorial assessment; test it on your own workload.
Qwen3-VL-30B-A3B-Thinking is a reasoning-focused multimodal MoE variant intended to combine text generation with image and video understanding. Its model card reports strengths in STEM and math reasoning, OCR, spatial grounding, long-document and long-video comprehension, GUI operation, and image-to-code generation. OpenRouter lists text and image inputs with text output, mandatory reasoning, a long context window, tool use, and structured outputs. These are source-reported capabilities; the supplied evidence does not provide independently measured results.[1][2]
Strengths
- The Thinking variant is explicitly positioned for multi-step STEM, mathematics, causal analysis, and evidence-based answers.[1][2]
- Reported vision features include multilingual OCR, spatial and temporal grounding, visual recognition, GUI element understanding, and generation of diagram or web code from visual inputs.[2]
- The catalog lists a long context window, text-plus-image-to-text modality, tool or function calling support, structured outputs, and several standard decoding controls.[1]
Limitations
- The evidence is a catalog record plus vendor documentation; independently measured benchmark results were not supplied, so accuracy, robustness, latency, and total cost require workload validation.[1][2]
- The sources differ on input scope: the documentation describes video understanding, while OpenRouter lists only text and image inputs for this catalog entry.[1][2]
- OpenRouter marks reasoning as mandatory and lists no supported effort levels; the evidence does not show a non-reasoning mode or measured latency impact.[1]
Where to use it
- Consider for long-document visual QA and multilingual OCR, after confirming whether the chosen endpoint accepts only text and image inputs or also video.[1][2]
- Consider for STEM or math assistance that needs explicit reasoning grounded in visual inputs.[1][2]
- Consider for prototypes involving GUI operation or image-to-interface-code generation, with human review because the pack supplies reported capabilities rather than measured results.[2]
Sources
- [1] OpenRouter catalog · Qwen: Qwen3 VL 30B A3B ThinkingSources checked: Nov 13, 2026
- [2] Model documentation · Qwen: Qwen3 VL 30B A3B ThinkingSources checked: Nov 13, 2026
About this model
Description below is the original text supplied by OpenRouter, shown in the source language.
Qwen3-VL-30B-A3B-Thinking is a multimodal model that unifies strong text generation with visual understanding for images and videos. Its Thinking variant enhances reasoning in STEM, math, and complex tasks. It excels...
Specifications
- Model ID
- qwen/qwen3-vl-30b-a3b-thinking
- Creator
- qwen
- Added to OpenRouter
- 06.10.2025
- Context length
- 262,144
- Modality
- text+image->text
- Input modalities
- text, image
- Output modalities
- text
- Token pricing
- USD per 1M tokens: in $0.290 / out $1
- Knowledge cutoff
- 31.03.2025
- Expiration date
- 15.11.2026
Tasks and capabilities
A grade appears only where the source supplies a matching measurement for that task, the model is active and the cohort holds at least 10 comparable models. The capability column lists facts the source declares — supported formats, reasoning and tool parameters — which are not measured quality.
| Task | Grade | Source-declared capability |
|---|---|---|
| General quality | No measurement | reasoning: mandatory · parameters: include_reasoning, reasoning |
| Coding | No measurement | — |
| Agents | No measurement | parameters: structured_outputs, tool_choice, tools |
| Writing | No measurement | output: text |
| Reasoning | No measurement | reasoning: mandatory · parameters: include_reasoning, reasoning |
| Document analysis | No measurement | — |
| Vision | No measurement | input: image |
| Image generation | No measurement | — |
| Audio | No measurement | — |
| Video | No measurement | — |
| Embeddings | No measurement | — |
Evaluations
No evaluations are supplied for this model.
How grades work
Each letter grade compares a model only against other models scored by the same source and metric in the current snapshot. S means the top 10%, A the next 20%, B the next 30%, C the rest. Models without a published measurement are shown as not rated — a supported format alone never implies measured quality.
- S — S — top 10% of the cohort
- A — A — next 20% of the cohort
- B — B — next 30% of the cohort
- C — C — below the B band
Providers
Optional live provider statistics from OpenRouter. Missing values are unknown, not zero.
| Provider | Context length | Price in /1M | Price out /1M | Latency p50 (30m, s) | Throughput p50 (30m, tok/s) | Uptime (30m) |
|---|---|---|---|---|---|---|
| SiliconFlowfp8 | 262,144 | $0.290 | $1 | — | — | 100% |
Provider data retrieved: 13.11.2026
Related articles
No published articles identify this model yet.