Data sourced primarily from official releases (GitHub, Hugging Face, papers), then benchmark leaderboards, then third-party evaluators. Learn about our data methodology

Qwen3-235B-A22B-Thinking

Model basics

Reasoning traces

Supported

Context length

256K tokens

Max output length

16384 tokens

Model type

推理大模型

Release date

2025-07-30

Model file size

31.17GB

MoE architecture

Yes

Total params / Active params

305.0B / 33B

Knowledge cutoff

No data

Inference modes

思考模式（Thinking Mode）

Qwen3-235B-A22B-Thinking

Open source & experience

Code license

Apache 2.0

Weights license

Apache 2.0- 免费商用授权

GitHub repo

https://github.com/QwenLM/Qwen3

Hugging Face

https://huggingface.co/Qwen/Qwen3-235B-A22B-Thinking-2507

Live demo

https://chat.qwen.ai/

Qwen3-235B-A22B-Thinking

Official resources

Paper

Qwen3: Think Deeper, Act Faster

DataLearnerAI blog

No blog post yet

Qwen3-235B-A22B-Thinking

API details

API speed

3/5

💡Default unit: $/1M tokens. If vendors use other units, follow their published pricing.

Standard pricingStandard

Modality	Input	Output
Text	$0.2	$2.4

Qwen3-235B-A22B-Thinking

Benchmark Scores

综合评估

4 evaluations

Benchmark / mode

Score

Rank/total

MMLU ProThinking

84.40

22 / 111

GPQA DiamondThinking

81.10

37 / 146

LiveBenchThinking

63.42

39 / 52

HLEThinking

18.20

48 / 91

编程与软件工程

1 evaluations

Benchmark / mode

Score

Rank/total

LiveCodeBenchThinking

74.10

25 / 102

数学推理

3 evaluations

Benchmark / mode

Score

Rank/total

AIME2025Thinking

92.30

31 / 104

IMO-ProofBenchThinking

33.30

6 / 16

IMO-ProofBench AdvancedThinking

5.20

5 / 8

写作和创作

1 evaluations

Benchmark / mode

Score

Rank/total

Creative WritingThinking

86.10

5 / 22

查看评测深度分析与其他模型对比

Qwen3-235B-A22B-Thinking

Publisher

阿里巴巴

View publisher details

Qwen3-235B-A22B-Thinking-2507

Model Overview

Qwen3-235B-A22B-Thinking-2507是阿里开源的Qwen3-30B-A3B模型升级版本，前者是2025年4月28日阿里开源的混合思考模式的模型（即一个模型支持推理和非推理模式），三个月后阿里发布了新版本，将其拆分为只支持推理和不支持推理的两个版本。Qwen3-235B-A22B-Thinking-2507是其中仅支持推理模式的版本，但是效果提升明显。

Qwen3-235B-A22B-Thinking-2507模型与其它模型的评测结果对比：

https://www.datalearner.com/ai-models/ai-benchmarks-tests/compare-result?benchmarkInputString=16,32,42,44&modelInputString=642,641,562,567,566

DataLearner 官方微信

欢迎关注 DataLearner 官方微信，获得最新 AI 技术推送