Qwen3-Coder provides a code-focused model in the Qwen family of large language models.
Qwen3
The Qwen3 model series was developed by the Qwen team at Alibaba Cloud.
About Qwen3
Qwen3 is a large language model series developed by the Qwen team at Alibaba Cloud. The repository shares model weights in dense and Mixture-of-Experts architectures, including sizes 0.6B, 1.7B, 4B, 8B, 14B, 32B, 30B-A3B and 235B-A22B, with updated 2507 Instruct and Thinking variants in 4B, 30B-A3B and 235B-A22B.
The repository includes documentation and examples for inference with Transformers, running locally on CPU and GPU, large-scale deployment, quantization, and post-training. Models support a thinking mode for reasoning, math and coding and a non-thinking mode for general chat, with integration with external tools for agent use. The 2507 variants list 256K-token context handling extendable to 1 million tokens.
Key features
- Dense and MoE weights from 0.6B to 235B-A22B
- Thinking mode for reasoning, math and coding tasks
- Non-thinking mode for general-purpose chat
- Local CPU and GPU use via llama.cpp, Ollama, LM Studio
- Deployment guidance for SGLang, vLLM and TGI
- Quantization with GPTQ, AWQ and GGUF guidance
Best for: Developers and researchers who need publicly available model weights for inference, local experiments and deployment.
Good for
- Publicly available weights in dense and MoE architectures
- Documentation covers inference, local use, deployment and quantization
- Offers thinking and non-thinking modes for different tasks
Qwen3 at a glance
- Website
- github.com/QwenLM/Qwen3
- Category
- AI & Machine Learning › Models & ML Research
- Type
- Open-source project
- Pricing
- Not confirmed. Check the site
- Open source
- Not stated
- Language
- English
- Last checked
- Oct 6, 2026 · Online when checked
- Listed
- Oct 5, 2026
Qwen3 alternatives
Tools and sites you could use instead, ranked by how closely they match.
WizardLM collects language models developed with Evol-Instruct, including WizardCoder and WizardMath.
Janus combines multimodal understanding and image generation models in a unified model series.
Open source
24 alternatives
Point·E is a research project that generates 3D point clouds using diffusion models.
Open source
24 alternatives
MobiLlama is a small language model designed for edge devices, according to its repository description.
Open source
24 alternatives
This repository accompanies research on improving language model factuality and reasoning through multi-agent debate.
Popular in Models & ML Research
Well-known listings from the same topic.
Falcon LLMfalconllm.tii.aeERNIEernie.baidu.comTokenVersetoken-verse.github.iowTAObittensor.comAlpacacrfm.stanford.edu/2023/03/13/alpaca.htmlSmolVLAhuggingface.co/blog/smolvla
Qwen3 FAQ
What sizes are available?
Listed sizes include 0.6B, 1.7B, 4B, 8B, 14B, 32B, 30B-A3B and 235B-A22B, with 2507 variants in 4B, 30B-A3B and 235B-A22B.
What is the difference between thinking and non-thinking modes?
Thinking mode is for complex logical reasoning, math and coding, while non-thinking mode is for efficient, general-purpose chat.
How can Qwen3 be run locally?
The documentation provides instructions for running on CPU and GPU with frameworks like llama.cpp, Ollama and LM Studio.
Is something wrong or out of date?
Tell us and an editor will check it.
Own this website?
Claim this listing
Claim the listing and suggest updates yourself. An editor checks every claim first.