Mistral Unveils Medium 3 Large Model: High Cost-Effectiveness Meets Easy Deployment

Mistral AI has launched the Mistral Medium 3 language model, which combines high cost-effectiveness with ease of deployment. The model delivers performance comparable to that of large-scale models while costing 8 times less. It only costs $0.4 per million input tokens and $2 for output. It exhibits excellent performance in specialized fields such as coding and multimodal understanding. It supports hybrid or on-premises deployment, customized post-training, and integration with enterprise systems.