Qwen3-4B-Thinking-2507

Name: Qwen3-4B-Thinking-2507
Availability: InStock
Author: 阿里巴巴

聊天大模型

Release date: 2025-08-06更新于: 2025-08-07 10:45:36915

Live demo GitHub Hugging Face Compare

Parameters

40.0亿

Context length

256K

Chinese support

Supported

Reasoning ability

Qwen3-4B-Thinking-2507 is an AI model published by 阿里巴巴, released on 2025-08-06, for 聊天大模型, with 40.0B parameters, and 256K tokens context length, requiring about 8.05GB storage, under the Apache 2.0 license.

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

Qwen3-4B-Thinking-2507

Model basics

Reasoning traces

Supported

Thinking modes

Thinking modes not supported

Context length

256K tokens

Max output length

16384 tokens

Model type

聊天大模型

Release date

2025-08-06

Model file size

8.05GB

MoE architecture

Total params / Active params

40.0B / N/A

Knowledge cutoff

No data

Qwen3-4B-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/Qwen3-4B-Thinking-2507

Live demo

https://chat.qwen.ai/

Qwen3-4B-Thinking-2507

Official resources

Paper

Qwen3: Think Deeper, Act Faster

DataLearnerAI blog

No blog post yet

Qwen3-4B-Thinking-2507

API details

API speed

4/5

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

Standard pricingStandard

Modality	Input	Output
Text	$0.11	$1.26

Qwen3-4B-Thinking-2507

Benchmark Results

Qwen3-4B-Thinking-2507 currently shows benchmark results led by AIME2025 (55 / 106, score 81.30), LiveCodeBench (73 / 109, score 55.20), GPQA Diamond (116 / 166, score 65.80). This page also consolidates core specs, context limits, and API pricing so you can evaluate the model from benchmark results and deployment constraints together.