NEWHow AI models draw the world map · Gecko Tests
Beta
Home/Models/Qwen3 4B Thinking 2507
Alibaba logo

Qwen3 4B Thinking 2507

by Alibaba · Released Aug 2025

Open Source
48.4
avg score
Rank #156
Compare
Better than 50% of all models
Context
N/A
Input $/1M
n/a
Output $/1M
n/a
Type
text-generation
License
Open Source
Benchmarks
6 tested
Data as of
About

Qwen text generation model. 1217K downloads on HuggingFace.

Tested on 6 benchmarks · BenchGecko score 48.4. Top scores: OpenCompass — IFEval (88.5%), OpenCompass — AIME2025 (80.0%), OpenCompass — MMLU-Pro (72.8%).

Capabilities
coding
51.6
#87 globally
math
80.0
#25 globally
knowledge
47.8
#144 globally
language
88.5
#19 globally
Benchmark Scores
Compare All
Tested on 6 benchmarks · Ranked across 4 categories
Score Distribution (all 312 models)
0255075100
▲ You are here
codingCompare coding →
OpenCompass — LiveCodeBenchV6

OpenCompass Live Code Bench v6. Fresh competitive programming problems to evaluate code generation without memorization.

51.6·
mathCompare math →
OpenCompass — AIME2025

OpenCompass evaluation on AIME 2025 problems. Tests mathematical reasoning on fresh competition problems.

80.0·
knowledgeCompare knowledge →
OpenCompass — MMLU-Pro

OpenCompass MMLU-Pro evaluation. Harder knowledge test with more answer choices.

72.8·
OpenCompass — GPQA-Diamond

OpenCompass evaluation of GPQA Diamond. PhD-level science questions from the hardest subset.

64.7·
OpenCompass — HLE

OpenCompass evaluation of Humanitys Last Exam. Expert-level cross-discipline knowledge test.

6.0·
Quick compare:Microsoft logovs Phi 4 Mini InstructMicrosoft logovs Phi 4 Mini InstructOpenAI logovs GPT-5 Pro
Excellent (85+) Good (70-85) Average (50-70) Below (<50)
Model Family · Alibaba Qwen 3
Apr 2025
n/an/actx
Apr 2025
n/an/actx
Apr 2025
n/an/actx
Apr 2025
n/an/actx
May 2025
n/an/actx
Jul 2025
n/an/actx
May 2025
n/an/actx
Apr 2025
n/an/actx
May 2025
n/an/actx
Apr 2025
n/an/actx
Aug 2025
20.3
n/an/actx6 benchmarks
Aug 2025
n/an/actx
Aug 2025
48.4+28.1
n/an/actx6 benchmarks
May 2025
n/an/actx
Apr 2025
n/an/actx
Jan 2026
n/an/actx
Jan 2026
n/an/actx
Jul 2025
n/an/actx
Feb 2026
n/an/actx
Jun 2025
n/an/actx
Jun 2025
n/an/actx
Jun 2025
n/an/actx
Jan 2026
n/an/actx
May 2025
n/an/actx
Jun 2025
n/an/actx
Jan 2026
n/an/actx
Jan 2026
n/an/actx
Jan 2026
n/an/actx
Jan 2026
n/an/actx
Oct 2025
n/an/actx
Oct 2025
n/an/actx
Oct 2025
n/an/actx
Oct 2025
n/an/actx
Oct 2025
n/an/actx
Jan 2026
n/an/actx
Jan 2026
n/an/actx
See the full Qwen 3 family →
Similar Models
Microsoft logo
Phi 4 Mini Instruct
Microsoft
48.4n/a
Microsoft logo
Phi 4 Mini Instruct
Microsoft
48.4$0.08/1M
OpenAI logo
GPT-5 Pro
OpenAI
48.7$15.00/1M
Links
Info
AlibabaPricing explorerDevelopers · API
Research
Technical Report
Documentation
API DocsPlayground
Community
@Alibaba
Source Code
GitHubHugging Face
BenchGecko API
qwen-qwen3-4b-thinking-2507
Specifications
  • Typetext-generation
  • ContextN/A
  • ReleasedAug 2025
  • LicenseOpen Source
  • StatusActive
Available On
Alibaba logoAlibaban/a
Categories
codingmathknowledgelanguage
Learn More
transformeropen weightstokens
Share & Export
Tweet
Related Models
Microsoft logoPhi 4 Mini Instruct
48.4
Microsoft logoPhi 4 Mini Instruct
48.4
OpenAI logoGPT-5 Pro
48.7
Meta logoLlama 3.3 70B Instruct
47.9
Qwen3 4B Thinking 2507 is an open-source text-generation AI model by Alibaba, released in August 2025. It has an average benchmark score of 48.4.

Related Models

Phi 4 Mini Instruct · MicrosoftPhi 4 Mini Instruct · MicrosoftGPT-5 Pro · OpenAILlama 3.3 70B Instruct · MetaGLM 5.3 Flash · z-ai

Benchmarks

OpenCompass — IFEvalOpenCompass — AIME2025OpenCompass — MMLU-ProOpenCompass — GPQA-DiamondOpenCompass — LiveCodeBenchV6

Related Pages

Alibaba · ProviderAlibaba · EconomyAll ModelsCompare ModelsPricingDevelopers · API

Key facts · as of 2026-04-09

  • Qwen3 4B Thinking 2507 by Alibaba. BenchGecko score 48.4, rank 154 of 312 scored models (normalized average of public benchmark scores).
  • List price n/a input · n/a output per 1M tokens (as of 2026-04-09).

How to cite · data as of 2026-04-09

Qwen3 4B Thinking 2507 · benchmarks, pricing and providers. BenchGecko, data as of 2026-04-09. https://benchgecko.ai/model/qwen-qwen3-4b-thinking-2507

Credit "Source: BenchGecko" with a link. Prices per provider and Gecko Tests are BenchGecko data (CC BY 4.0); benchmark scores keep their original source, listed in the JSON. JSON · llms.txt · MCP