Mistral has released the weights of Mistral Large 4, nicknamed Le Chonk, a 1-trillion-parameter multimodal model aimed at coding and defensive security work. Weights are the downloadable model files, so you can run the model on your own servers instead of calling Mistral’s.
The license has one commercial catch
Mistral calls the license a modified MIT license. You can use, modify, fine-tune and commercialize the model. The exception: a business that operates a Model-as-a-Service offering, meaning it sells access to the model itself, needs a separate commercial license from Mistral first.
API prices
Through Mistral’s API, Mistral lists:
- Input: $0.70 per million tokens
- Cached input (a repeated prompt prefix the provider has stored): $0.07 per million tokens
- Output: $2.10 per million tokens
The model is also listed on Mistral Studio, OpenRouter, Microsoft Foundry and NVIDIA-hosted endpoints. For self-hosting, Mistral points to Hugging Face downloads and the vLLM and SGLang serving tools. A model this size needs serious GPU hardware, so most solo builders will use a hosted endpoint.
Reasoning modes and benchmarks
Large 4 has three reasoning modes: low for simple tasks, medium for everyday tasks, and high, the default, for the hardest problems.
Mistral reports 19.7% on FrontierCode 2.0 and 41% on Terminal-Bench 4.0, and says the model leads open-weight models at coding. Results on your own codebase may differ.
End of article