Qwen3 VL 235B A22B Thinking
Qwen · released Sep 22, 2025 · Qwen/Qwen3-VL-235B-A22B-Thinking
- Type
- Open weightsApache License 2.0
- Params
- 236B
- Context
- 131K
22B active per word · about 98K words of context
Our take
Written Sep 30, 2026Qwen3 VL 235B A22B Thinking is a downloadable model that takes text and images and returns text, with a licence that allows commercial use, changes and redistribution (Apache License 2.0). Its measured quality sits mid-field on the boards we hold, so it is a general-purpose pick rather than a leader.
Use it for general chat, image questions and document work where you want a model you can download and run yourself under permissive terms. It accepts text and images in one request, and its active footprint per token is far below its total size. Skip it if you need a top-of-board result on coding, maths or creative writing, where it sits mid-field.
The case for it
- You can download it and run it yourself, and the licence allows commercial use, changes and redistribution (Apache License 2.0).
- Images go into the same request as the text, so a screenshot or diagram does not have to be described in words first.
- 235.7 billion parameters in total, of which about 22 billion are used per token, so memory in use is far below its total size.
The case against it
- Mid-field on the boards we hold: 100th of 168 on Arena Text (overall) via Thinking as of 25 Sep 2026, and 92nd of 168 on Arena Coding on the same date.
- Those placings are human preference rankings, not correctness scores, so they say which answers people preferred rather than whether the work was right.
- Released 2025-09-22, so it competes with newer releases; that is a date, not a measured quality gap.
How good is it?
EverydayGeneral questions and everyday reasoning
Arena Text (overall)100th of 168 · 1395
CodingWriting and fixing code on its own
Arena Coding92nd of 168 · 1452
Arena Coding is the only board that has scored it for this.
AgenticPlanning, calling tools, staying on task
Not yet scored on Arena Agent.
WritingDrafting and rewriting prose
Arena Creative Writing106th of 168 · 1339
Arena Creative Writing is the only board that has scored it for this.
These tests check whether a model follows instructions — a precondition for all the work above, but not a measure of how well that work is done.
Every published score for this model6 scoresEvery figure we hold, from 6 boards, with who ran it and a link to the source — including the boards no rating above is built on.
Can you run it yourself?
GeForce RTX 4090 · 24 GB
Too large for this card. The weights do not fit even with part of them offloaded to system memory.
Apple M1 Pro (16-core GPU) · 32 GB
Too large for this card. The weights do not fit even with part of them offloaded to system memory.
Comfortable fit
Apple M3 Ultra (80-core GPU) · 512 GB
Room to spare. 229.3 GB spare means a 10% error in the size would not change the answer.
Memory use by level
Against a 24 GB card.
What is quantisation? →This model on every device we track71 devicesThe Q4 build most people download, on each device: what the weights come to, how much context the memory leaves, and whether it runs. Smallest device that runs it first.
Check against your own machine → · Where to rent it hosted →
Or rent it from someone else
Prices checked between 4 hours and 10 hours ago — each listing carries its own date.
Alibaba Cloud, through OpenRouter
Cheapest of the 2 listings we can compare like for like — at 131K of context, out of 3 in the table below. One cheaper row there is outside that comparison: a different quantisation.
- per 1M tokens
- $0.40 in / $4.00 out
- Context served
- 131K
- Throughput
- ~57 tok/s
| Provider | In / out per 1M tokens | Context | Throughput | Trains on prompts | Logs prompts | Zero retention |
|---|---|---|---|---|---|---|
| Novita AIbf16Direct and through OpenRouter | $0.98 / $3.95checked 4 hours ago | 131K33K max reply through OpenRouter | 43 tok/sthrough OpenRouter | DirectUnknownThrough OpenRouterNo | DirectUnknownThrough OpenRouterNo | DirectUnknownThrough OpenRouterConfirmed |
| OpenRouterOpenRouter's own listing | $0.40 / $4.00checked 4 hours ago | 131K | not measured | Unknown | Unknown | Unknown |
| Alibaba CloudThrough OpenRouter | $0.40 / $4.00checked 10 hours ago | 131K33K max reply | 57 tok/s | No | Yesunknown period | Unknown |
Novita AI
OpenRouterAcross the 3 listings we hold: 2 say they do not train on prompts (1 of them only through OpenRouter), 0 say they do and 1 does not say. 1 appears in the zero-retention registry we check (only through OpenRouter); the rest are unknown to us.
What each host's API supports
From the parameter list each endpoint publishes. Streaming is omitted: nothing we hold reports it, for any model.
| Provider | Tool calling | JSON output | Strict schema |
|---|---|---|---|
| Novita AIbf16Direct and through OpenRouter | ✓ | ✗ | ✗ |
| OpenRouterOpenRouter's own listing | ✓ | ✓ | ✓ |
| Alibaba CloudThrough OpenRouter | ✓ | ✓ | ✓ |
Tool calling: 3 of 3 listings say yes. JSON output: 2 of 3 listings say yes, 1 says no. Strict schema: 2 of 3 listings say yes, 1 says no.
When we formed this view
Recent changes
What moved
first indexed by our pipelineEach date is the day we first saw the change, or the day the maker announced it.
What we do not know about this model yet
- We hold no measured file for it, so all 3 sizes on this page are calculated from the parameter count.
- Nothing we hold says whether an endpoint streams, so we do not show it either way.
- 1 of 3 listings does not say whether it trains on prompts, and 1 answers only through OpenRouter, not for its own listing.
- We hold no cached-input rate for any of its listings.
- We hold no batch or off-peak rate for any of its listings.
- We hold a decode speed for it, but no prompt-processing (prefill) figure, so how long the input side of a job takes is unknown to us.
Licence and identifiers
What the licence allowsApache License 2.0, what it allows commercially, and the identifiers you need to pull this model — its Hugging Face repo, our slug and a machine-readable card.
Licence
Apache License 2.0
Fully permissive: commercial use, redistribution, and derivatives allowed. Requires attribution and a copy of the license. Includes an express patent grant.
Identifiers
- Hugging Face
- Qwen/Qwen3-VL-235B-A22B-Thinking
- Architecture
- Mixture of experts
- Takes in, gives back
- Text and images in, text out
- Catalogue slug
- qwen-qwen3-vl-235b-a22b-thinking