Mistral AI published the open weights for Mistral Large 4 on October 29, after about three weeks of public preview. According to Techstrong.ai, the multimodal model has roughly 1.05 trillion total parameters with about 52 billion active per forward pass, a one-million-token context window and support for more than 160 languages. The weights are on Hugging Face, and the model is also available through Mistral's API, OpenRouter, Nvidia-hosted endpoints and Microsoft Foundry.
The model ships under the Mistral Ecosystem Open-Weight License, a modified MIT license that allows use, modification, fine-tuning and commercial deployment. There is one condition: companies that offer the model as a service to third parties need a separate commercial license from Mistral. Mistral also halved API pricing, to $0.70 per million input tokens and $2.10 per million output tokens.
Artificial Analysis scored the model 26 on its Intelligence Index in high reasoning mode. That makes it the top open-weight model built outside China, though it still trails DeepSeek V4.1 Flash and GLM-5.3. A specialized cybersecurity variant is available only to vetted design partners.