Mistral releases open weights for Large 4

Mistral has released the weights of Mistral Large 4, also known as Le Chonk, following its public preview. The 1-trillion-parameter multimodal model was trained from scratch on European infrastructure and supports deployment on organizations’ own systems. Mistral says additional reinforcement learning has improved coding, defensive cybersecurity and reasoning efficiency.

On the Artificial Analysis Coding Agent Index, the model scores 48%, effectively matching GLM-5.3 while using 47% fewer tokens per task on average. It also leads open-weight models on FrontierCode and Terminal-Bench, according to Mistral. Developers can choose low, medium or high reasoning modes to balance performance and inference costs.

For defensive security, the model scores 51% on the Artificial Analysis Cyber Index, which tests vulnerability discovery, reproduction and patching. Mistral reports an average refusal rate of 96.6% on malicious cyber requests. The company nevertheless acknowledges potential for the model to assist real attacks and is updating its guardrails. A specialized cybersecurity variant is being offered to vetted enterprise partners under contractual safeguards and authorized-testing restrictions.

The weights are available through Hugging Face, with hosted access through the Mistral API, OpenRouter and Microsoft Foundry, among other options. Organizations can customize the model using Mistral Forge. Its Mistral Ecosystem Open-Weight license permits use, modification, fine-tuning and commercialization, but businesses operating Model-as-a-Service offerings must obtain a separate commercial license before commercial use.