AI BriefingMedium becomes the new Large
Key point
Mistral AI has unveiled Mistral Medium 3, achieving both high performance and 8x lower cost.
Details
Mistral AI announced Mistral Medium 3, a model that maximizes efficiency and usability. This model maintains state-of-the-art (SOTA) performance while achieving 8x lower cost compared to existing models, and has significantly improved deployment convenience for rapid enterprise adoption.
In terms of performance, Mistral Medium 3 records over 90% of Claude Sonnet 3.7's performance, surpassing major models such as Llama 4 Maverick and Cohere Command A. It shows particularly outstanding capability in coding and STEM tasks.
Key features are as follows:
- Overwhelming cost-efficiency: Priced at $0.4 per 1M input tokens and $2 per 1M output tokens, it is more economical than DeepSeek v3.
- Flexible deployment: Can be operated in any cloud environment, as well as in self-hosted environments with 4 or more GPUs.
- Enterprise features: Supports hybrid, on-premise, and VPC deployment, along with custom post-training.
This summary was generated automatically by AI. Check the original for the author's claims and context. Copyright belongs to the original author.
Our guide explains how the AI works. Report summary errors, attribution issues, or removal requests via Contact.