Mistral AI ✓

New Apache 2.0 open license for the whole family, instead of the older research-style license

Switch to a Mixture-of-Experts architecture for the flagship (more total params, fewer active, better efficiency)

Multimodal by default, with built-in image understanding instead of separate vision models

Context window doubled to 256k tokens for the flagship and small models

Expanded small model lineup (Ministral 3B/8B/14B, base/instruct/reasoning) tuned for edge and reasoning use cases, with much lower API prices overall

Mistral Large 2 has a 128k context window and supports dozens of languages including French, German, Spanish, Italian, Portuguese, Arabic, Hindi, Russian, Chinese, Japanese, and Korean, along with 80+ coding languages including Python, Java, C, C++, JavaScript, and Bash.

Mistral Large 2 is designed for single-node inference with long-context applications in mind – its size of 123 billion parameters allows it to run at large throughput on a single node. We are releasing Mistral Large 2 under the Mistral Research License, that allows usage and modification for research and non-commercial usages. For commercial usage of Mistral Large 2 requiring self-deployment, a Mistral Commercial License must be acquired by contacting us.

Mistral Large 2 sets a new frontier in terms of performance / cost of serving on evaluation metrics. In particular, on MMLU, the pretrained version achieves an accuracy of 84.0%, and sets a new point on the performance/cost Pareto front of open models.

Mistral Large 2 has a 128k context window and supports dozens of languages including French, German, Spanish, Italian, Portuguese, Arabic, Hindi, Russian, Chinese, Japanese, and Korean, along with 80+ coding languages including Python, Java, C, C++, JavaScript, and Bash.

Mistral Large 2 is designed for single-node inference with long-context applications in mind – its size of 123 billion parameters allows it to run at large throughput on a single node. We are releasing Mistral Large 2 under the Mistral Research License, that allows usage and modification for research and non-commercial usages. For commercial usage of Mistral Large 2 requiring self-deployment, a Mistral Commercial License must be acquired by contacting us.

Mistral Large 2 sets a new frontier in terms of performance / cost of serving on evaluation metrics. In particular, on MMLU, the pretrained version achieves an accuracy of 84.0%, and sets a new point on the performance/cost Pareto front of open models.