QwenLM /Qwen2.5
Qwen3 is the large language model series developed by Qwen team, Alibaba Cloud.
★ 27,732⑂ 2,105👁 167⑃ 135● 1,243▤ 0
More from this author
+1
Daily
+27
Weekly
+61
Monthly
27,732
stars
2,105
forks
167
watchers
135
prs
1,243
issues
0
releases
README
Qwen3
💜 Qwen Chat   |   🤗 Hugging Face   |   🤖 ModelScope   |    📑 Paper    |    📑 Blog    |   📖 Documentation
🖥️ Demo   |   💬 WeChat (微信)   |   🫨 Discord  
Visit our Hugging Face or ModelScope organization (click links above), search checkpoints with names starting with Qwen3- or visit the Qwen3 collection, and you will find all you need! Enjoy!
To learn more about Qwen3, feel free to read our documentation \[EN|ZH\]. Our documentation consists of the following sections:
- Quickstart: the basic usages and demonstrations;
- Inference: the guidance for the inference with Transformers, including batch inference, streaming, etc.;
- Run Locally: the instructions for running LLM locally on CPU and GPU, with frameworks like llama.cpp, Ollama, and LM Studio;
- Deployment: the demonstration of how to deploy Qwen for large-scale inference with frameworks like SGLang, vLLM, TGI, etc.;
- Quantization: the practice of quantizing LLMs with GPTQ, AWQ, as well as the guidance for how to make high-quality quantized GGUF files;
- Training: the instructions for post-training, including SFT and RLHF (TODO) with frameworks like Axolotl, LLaMA-Factory, etc.
- Framework: the usage of Qwen with frameworks for application, e.g., RAG, Agent, etc.
Introduction
Qwen3-2507
Over the past three months, we continued to explore the potential of the Qwen3 families and we are excited to introduce the updated Qwen3-2507 in two variants, Qwen3-Instruct-2507 and Qwen3-Thinking-2507, and three sizes, 235B-A22B, 30B-A3B, and 4B.
Qwen3-Instruct-2507 is the updated version of the previous Qwen3 non-thinking mode, featuring the following key enhancements:
- Significant improvements in general capabilities, including instruction following, logical reasoning, text comprehension, mathematics, science, coding and tool usage.
- Substantial gains in long-tail knowledge coverage across multiple languages.
- Markedly better alignment with user preferences in subjective and open-ended tasks, enabling more helpful responses and higher-quality text generation.
- Enhanced capabilities in 256K-token long-context understanding, extendable up to 1 million tokens.
Qwen3-Thinking-2507 is the continuation of Qwen3 thinking model, with improved quality and depth of reasoning, featuring the following key enhancements:
- Significantly improved performance on reasoning tasks, including logical reasoning, mathematics, science, coding, and academic benchmarks that typically require human expertise — achieving state-of-the-art results among open-weight thinking models.
- Markedly better general capabilities, such as instruction following, tool usage, text generation, and alignment with human preferences.
- Enhanced 256K long-context understanding capabilities, extendable up to 1 million tokens.
Previous Qwen3 Release
Qwen3 (aka Qwen3-2504)
Top issues
#1568vllm启动QwQ以及Qwen3-32B 模型,使用functioncall功能时,流式调用接口时,argument变量返回始终只能得到一个值,而非流式调用时一切正常24 comments · 2025-07-24#925[Badcase]: 用ollama运行 qwen 2.5. 结尾回复会一直重复19 comments · 2024-09-19#1384[Bug]: vllm部署qwen3-30b-a3b输出中文乱码 [Badcase]: Qwen3-30B-A3B输出重复停不下来15 comments · 2025-05-09#1826Chat template breaks KV-cache reuse when enable_thinking=false9 comments · 2026-03-03#1638Hugging Face Qwen3 demo space broken11 comments · 2025-09-07#1341[Badcase]: qwen3复述用户输入的网址出错9 comments · 2025-04-30#684有无量化到24gB以内的?支持 mermaid diagrams and UML diagrams9 comments · 2024-06-20#997[Bug]: 文档问答会忽略部分数据,比如证书号是12345 回答的是23457 comments · 2024-09-29
Discussions
#1931Request for technical feedback on an independent language model (CatAI / Purr 1.0)0 comments · 2026-11-14#1930Qwen3.5-4B + LoRA — Advice on reliable multi-turn constraint tracking and entity grounding1 comments · 2026-11-10#1929Running Qwen locally on Apple Silicon as an agent backend with Rapid-MLX0 comments · 2026-11-06#1928Multidirectional Resonant Limiting (MRL): In-place <=5% path-gauge GGUF merging of Qwen2.5-1.5B & Qwen2.5-Coder-1.5B in Termux (0.0000% super-block loss)0 comments · 2026-11-05#1927Pretrained Qwen models for real-time news trading: NQ, ES and YM futures1 comments · 2026-11-04#1926用 Qwen3-0.6B 查看本地权重:张量结构与分布的同屏检查0 comments · 2026-10-31#1924OpenMontage Drama Studio: a self-hosted AI short-drama studio using Qwen (DashScope) and Qwen-Image0 comments · 2026-10-28#1923[Experiment Report / 实验报告] Testing Qwen3-4B in Physical Sandboxes: Why a 4B LLM Should Be a Language Layer, Not a Final State Machine0 comments · 2026-10-23