DataLearner logoDataLearnerAI
Latest AI Insights
Model Evaluations
Model Directory
Model Comparison
Resource Center
Tools

加载中...

DataLearner logoDataLearner AI

A knowledge platform focused on LLM benchmarking, datasets, and practical instruction with continuously updated capability maps.

产品

  • Leaderboards
  • 模型对比
  • Datasets

资源

  • Tutorials
  • Editorial
  • Tool directory

关于

  • 关于我们
  • 隐私政策
  • 数据收集方法
  • 联系我们

© 2026 DataLearner AI. DataLearner curates industry data and case studies so researchers, enterprises, and developers can rely on trustworthy intelligence.

隐私政策服务条款
Page navigation
目录
Model catalogQwen3-235B-A22B-Thinking-2507
QW

Qwen3-235B-A22B-Thinking-2507

Qwen3-235B-A22B-Thinking-2507

Release date: 2025-07-25更新于: 2025-07-27 23:27:051,193
Live demoGitHubHugging FaceCompare
Parameters
2350.0亿
Context length
256K
Chinese support
Supported
Reasoning ability

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-2507

Model basics

Reasoning traces
Supported
Thinking modes
Thinking modes not supported
Context length
256K tokens
Max output length
32768 tokens
Model type
推理大模型
Release date
2025-07-25
Model file size
470.77 GB
MoE architecture
Yes
Total params / Active params
2350.0B / 220B
Knowledge cutoff
No data
Qwen3-235B-A22B-Thinking-2507

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-2507

Official resources

Paper
Qwen3-235B-A22B-Instruct-2507
DataLearnerAI blog
阿里发布Qwen3小幅更新版本,放弃混合思考模式,发布全新的2个版本Qwen3-235B-A22B-2507模型,1/5的参数,性能直逼Kimi K2,推理模式版本评测结果接近o3
Qwen3-235B-A22B-Thinking-2507

API details

API speed
3/5
💡Default unit: $/1M tokens. If vendors use other units, follow their published pricing.
Standard pricingStandard
ModalityInputOutput
Text$0.7$8.4
Qwen3-235B-A22B-Thinking-2507

Benchmark Results

Thinking

综合评估

4 evaluations
Benchmark / mode
Score
Rank/total
MMLU Pro
default
84.40
25 / 114
GPQA Diamond
default
81.10
48 / 158
LiveBench
default
69.11
22 / 51
HLE
default
18.20
67 / 111

编程与软件工程

1 evaluations
Benchmark / mode
Score
Rank/total
LiveCodeBench
default
74.10
28 / 105

数学推理

1 evaluations
Benchmark / mode
Score
Rank/total
AIME2025
default
92.30
32 / 106

写作和创作

1 evaluations
Benchmark / mode
Score
Rank/total
Creative Writing
default
86.10
5 / 22
查看评测深度分析与其他模型对比
Qwen3-235B-A22B-Thinking-2507

Model variants & downloads

Variant nameVersion typeQuantizationModel sizeHuggingFace link
Qwen3-235B-A22B-Thinking-2507-FP8ℹ️InstructFP8236.45 GBDownload link
Qwen3-235B-A22B-Thinking-2507

Publisher

阿里巴巴
阿里巴巴
View publisher details
Qwen3-235B-A22B-Thinking-2507

Model Overview

阿里巴巴开源的Qwen3-235B-A22B模型的升级版本,最早的Qwen3-235B-A22B模型是在2025年4月28日随着Qwen3系列一起发布,当时是推理和非推理模式混合的架构模型,后来阿里发现这个模式不好,因此在2025年7月份发布了更新版的模型,即不支持推理模式的Qwen3-235B-A22B-2507和支持推理模式的Qwen3-235B-A22B-Thinking-2507。


Qwen3-235B-A22B-Thinking-2507最多可以支持80K的推理过程长度,最高支持32K的答案输出,是当前推理过程最长的模型之一!

DataLearner 官方微信

欢迎关注 DataLearner 官方微信,获得最新 AI 技术推送

DataLearner 官方微信二维码