Mistral releases open weights for Mistral Large 4, its 1-trillion-parameter flagship, on Hugging Face
On Oct 29 Mistral released the weights of Mistral Large 4 ("Le Chonk"), in API preview since early October. The multimodal MoE model has about 1T total and 49B active parameters and three reasoning modes. Mistral says three more weeks of RL lifted Terminal-Bench 4.0 from 28% to 41%, calling it the best open-weight coding model (vendor claim). A quantized version runs on four Nvidia B200/B300 GPUs.
Why it matters
One of the largest open-weight models from a European lab: developers can now self-host and fine-tune a 1T-parameter multimodal model. Mistral says smaller distilled versions, including a laptop-class one, will follow.
Sources5
- MistralOfficial · BlogMistral Large 4 (Le Chonk) is now open weightView original ·
- The Next WebMediaMistral releases open weights for its 1 trillion-parameter Large 4 modelView original ·
- Hugging FaceMediamistralai/Mistral-Large-4-1T-A52BView original ·
- @MistralAIXToday, we release the weights of a new and improved Le Chonk, advancing the open frontier for agentic coding and cybersecurity. 💻 Built to build: Le Chonk ranks #1 on @cognition’s FrontierCode among open models, and sets a new benchmark for Pareto efficiency on @artificialanalysis. 🛡 Built to defend: State-of-the-art in cyberdefense, put through the paces by the world's leading cyber vendors, including @CrowdStrike, @cloudflare, @rubrik, @zscaler, and others. Built for sovereignty: Make it yours. Download the weights on @huggingface , customize it with Mistral Forge, and deploy it wherever you need it. Read more from our team and customers 👇:View original ·
- @vllm_projectX🎉 Day-0 RL for @MistralAI’s Mistral Large 4 with @NovaSkyAI’s SkyRL and @vllm_project! On 8 @nvidia GB200 nodes, LoRA RL on the NVFP4 checkpoint brings AIME 2024 avg@8 from ~64% to ~75% in 75 training steps. vLLM generates rollouts, while FSDP2 or Megatron handles training.View original ·
Written by dalao.ai: the title, summary and why it matters are our own words. The original reports are the record; their publishers keep authorship.