News

NVIDIA and Mistral AI partner to co-develop frontier open-source AI models

Tuesday, March 17, 2026 at 12:20 AM

NVIDIA and Mistral AI have entered into a strategic partnership focused on the co-development of frontier open-source AI models. The collaboration aims to merge Mistral AI's model expertise with NVIDIA's hardware and software stack to accelerate model development and inference performance.

Context

On March 16, 2026, NVIDIA and Mistral AI announced a strategic partnership to co-develop frontier open-source AI models. As a founding member of the newly launched NVIDIA Nemotron Coalition, Mistral AI will integrate its specialized model architectures with NVIDIA's compute resources and synthetic-data generation pipelines. This collaboration aims to advance open-weight foundation models that rival proprietary offerings, specifically targeting enterprise-grade multimodal and reasoning capabilities. This news follows the successful December 2025 launch of the Mistral 3 family, which includes the Mistral Large 3 model featuring 675B total parameters and a 256K context window. Optimized for NVIDIA Blackwell GB200 NVL72 systems, these models achieved a 10x performance gain over the previous H200 generation. By standardizing on NVIDIA's NVFP4 quantization and Dynamo inference framework, the partnership secures a high-performance, open-source pipeline from data centers to edge devices like RTX PCs and Jetson modules.

Related Companies

Nvidia
Nvidia
NVDA
US