Back to the index
Alibaba Qwen · Qwen
No. 29

Qwen3-235B-A22B-Thinking-2507

Open-weight reasoning model with strong math and science.

ReasoningOpen weightsText
Field notes
Context window
262Ktokens
Max output
82Ktokens
Input price
$0.15/ 1M tokens
Output price
$1.50/ 1M tokens
Cached input
$0.11/ 1M tokens
Modalities
Text
Knowledge cutoff
Apr 2025
Released
Jul 2025
Parameters
235B
License
Apache-2.0
Approx. speed
—
Overview

What this model is

Qwen3-235B-A22B-Thinking-2507 is Alibaba's open-weight reasoning model, a 235B MoE with 22B active params and long chain-of-thought training. It is excellent at math, competition problems, and science QA, and is fully self-hostable under Apache-2.0. Its verbose reasoning traces and weaker tool-use reliability make it less dependable than the frontier for autonomous agentic coding.

Strengths

  • Best-in-class open-weight math reasoning
  • Apache-2.0 license, fully self-hostable
  • Strong GPQA and LiveCodeBench results
  • Efficient 22B active params keep inference tractable

Trade-offs

  • Verbose, slow chains of thought
  • Weaker tool-use for agentic coding
  • Text-only, no multimodal input
  • Below the frontier on long-horizon tasks

Best for

  • Advanced math and scientific reasoning
  • Competitive coding and algorithm work
  • Self-hosted reasoning deployments
  • Multilingual analytical tasks

Not ideal for

  • Latency-sensitive applications
  • Autonomous agentic coding
Capabilities

Capability profile

Normalized 0–100 scores, comparable across the whole catalog.

Reasoning82Coding72Math90Writing78Knowledge84Speed55Agentic66Vision5Multilingual88Long Context80
Reasoning
82
Coding
72
Math
90
Writing
78
Knowledge
84
Speed
55
Agentic
66
Vision
5
Multilingual
88
Long Context
80
Benchmarks

How it scores

Public benchmark results, with independent third-party results where available. Bars normalize percentages to 100 and Elo ratings to a 1500 ceiling.

BenchmarkResult
SWE-bench Proindependent
CodingSWE-bench Pro
21.4%
Aider Polyglotindependent
CodingAider
57.3%
MMLU-Pro
KnowledgeVendor-reported
84.4%
GPQA Diamondindependent
ReasoningIndependent aggregators
81.1%
AIME 2025
MathVendor-reported
92.3%
LiveCodeBenchindependent
CodingLiveCodeBench
70.7%
LiveBenchindependent
GeneralLiveBench
77.1%
Alternatives

Comparable models

Models in a roughly similar class — worth weighing against this one.

DeepSeek-R1-0528DeepSeek

Iconic open reasoning model — strong math at low cost, now a generation behind.

Qwen3.7-MaxAlibaba Qwen

Alibaba's proprietary agentic reasoning flagship with a 1M context.

GLM-5.2Zhipu AI

Open-weight coding flagship at a fraction of frontier cost.