Qwen3 235B A22B Thinking 2507
Qwen · released Jul 25, 2025 · Qwen/Qwen3-235B-A22B-Thinking-2507
- Type
- Open weightsApache License 2.0
- Params
- 235B
- Context
- 262K
22B active per word · about 197K words of context
Our take
Written Aug 2, 2026Qwen3 is a large downloadable model built for coding and long-document reasoning, with 22 billion of its 235 billion parameters active on each token. Its Apache licence and wide provider choice make it a flexible pick for technical workloads, though creative writing is a clear weak point.
Choose this for coding workloads where its measured coding score leads its other capabilities, or for long-document reasoning with a 262,144-token request limit. Pick it for budget-conscious Apache-licensed deployment, or where latency matters more than price. Skip it if you need image or audio input, if creative writing quality is central, or if real-time speed is critical.
The case for it
- Strongest measured skill is coding, with a 68-point lead over its weakest skill.
- Massive request limit for a fully downloadable model: 262,144 tokens.
- Dramatic mixture-of-experts efficiency — only about one-tenth of parameters active per token.
- Wide provider choice with a threefold price spread to optimise for cost or speed.
The case against it
- Creative writing is a clear weak point, 68 points below its coding score.
- Text-only; no image, audio or video input or output.
- Thinking mode adds latency, and fastest measured throughput is modest for real-time use.
How good is it?
IntelligencePuzzles, maths, exam questions
Arena Text (overall)76th of 143 · 1398.9
CodingWriting and fixing code on its own
Arena Coding77th of 143 · 1441.5
Arena Coding is the only board that has scored it for this.
AgenticPlanning, calling tools, staying on task
Nobody we watch has scored Qwen3 235B A22B Thinking 2507 for this. We would take the rating from Arena Agent (IPS).
WritingWe do not rate this
Two boards come close and neither tests writing: Arena Creative Writing asks people which of two replies they prefer, and LiveBench Language tests whether a model understood a passage. So we show where Qwen3 235B A22B Thinking 2507 placed and give it no mark out of five.
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, which is why they get no rating.
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?
- Fits in memory
- weights load entirely on the card
- Spills to system RAM
- some weights offload; much slower
- Too large
- will not load even with offload
- est
- size is calculated; the verdict could change by 10%
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. 229.7 GB spare means a 10% error in the size would not change the answer.
Memory use by level
Against a 24 GB card.
All 3 sizes here are calculated, not measured. We hold no measured file for this model, so each size comes from the parameter count and every verdict above inherits that uncertainty.
Check against your own machine → · All 71 devices, with every size →
Or rent it from someone else
Cheapest of 8 live listings. Picked at the widest standard context we hold, within one quantisation slice, so the numbers beside it are a price one host actually charges.
- per 1M tokens
- $0.23 in / $2.30 out
- Context served
- 262K
- Throughput
- Not measured
| Provider | In / out per 1M tokens | Context | Throughput | Trains on prompts | Logs prompts | Zero retention |
|---|---|---|---|---|---|---|
| Alibaba Cloud | $0.15 / $1.50 | 131K | 47 tok/s | No | Yesunknown period | Unknown |
| DeepInfrafp8 | $0.23 / $2.30 | 262K | not measured | Unknown | Unknown | Unknown |
| DeepInfrafp8 | $0.23 / $2.30 | 262K | 38 tok/s | No | No | Confirmed |
| OpenRouter | $0.23 / $2.30 | 262K | not measured | Unknown | Unknown | Unknown |
| Alibaba Cloudfp8 | $0.23 / $2.30 | 131K | 54 tok/s | No | Yesunknown period | Unknown |
| Novita AI | $0.30 / $3.00 | 131K | not measured | Unknown | Unknown | Unknown |
| Novita AIfp8 | $0.30 / $3.00 | 131K | 28 tok/s | No | No | Confirmed |
| Venice AIfp8 | $0.45 / $3.50 | 128K | 38 tok/s | No | No | Confirmed |
DeepInfra
Novita AIAcross the 8 listings we hold: 5 say they do not train on prompts, 0 say they do and 3 do not say. 3 appear in the zero-retention registry we check; the rest are unknown to us rather than confirmed either way.
What each host's API supports
From the parameter list each endpoint publishes. Streaming is omitted: nothing we hold reports it, for any model.
- ✓
- Supported
- ✗
- Not supported
- Not published
- host gave no parameter list
| Provider | Tool calling | JSON output | Strict schema |
|---|---|---|---|
| Alibaba Cloud | ✓ | ✓ | ✗ |
| DeepInfrafp8 | |||
| DeepInfrafp8 | ✗ | ✓ | ✗ |
| OpenRouter | ✓ | ✓ | ✗ |
| Alibaba Cloudfp8 | ✓ | ✓ | ✗ |
| Novita AI | |||
| Novita AIfp8 | ✓ | ✗ | ✗ |
| Venice AIfp8 | ✓ | ✗ | ✗ |
Tool calling: 5 of 8 listings say yes, 1 says no, 2 publish no parameter list. JSON output: 4 of 8 listings say yes, 2 say no, 2 publish no parameter list. Strict schema: 0 of 8 listings say yes, 6 say no, 2 publish no parameter list.
Models people weigh against Qwen3 235B A22B Thinking 2507
When we formed this view
Dates behind this page
Prices last checked 37h ago
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.
- 2 of 8 listings publish no parameter list, so what their API accepts is unknown to us.
- Nothing we hold says whether an endpoint streams, so we do not show it either way.
- 3 of 8 listings do not say whether they train on prompts.
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/Qwen3-235B-A22B-Thinking-2507
- Architecture
- Mixture of experts
- Modality record
- text->text
- Catalogue slug
- qwen-qwen3-235b-a22b-thinking-2507