Mistral releases the weights of Mistral Large 4, a 1-trillion-parameter model
TL;DR
- On 29 October Mistral released the weights of an improved Mistral Large 4, a 1-trillion-parameter Mixture-of-Experts model with 49 billion active parameters per token.
- Its API costs $0.70 per million input tokens and $2.10 per million output tokens, and a Model-as-a-Service business needs a separate licence.
- Mistral calls it the best open-weight model for coding, while Artificial Analysis scores it six points behind DeepSeek V4.1 Flash 1017 and GLM-5.3 overall.
Read the full story
Sign in with your email to read AI News. It’s free.