Mistral 与 NVIDIA 合作推进工程领域 AI,将发布物理与智能体评测基准
Mistral 与 NVIDIA 宣布合作开发 AI 原生工程工作流,把 Mistral Large 4 与 NVIDIA CUDA-X、Warp 及 Omniverse 结合,用于加速工程仿真并打通仿真、模型开发与设计之间的闭环。双方将推出面向物理模型的公开基准和开放数据集,并针对 CAE 和 EDA 场景中的智能体 AI 建立评测,考察工具调用、多步流程完成度以及执行时间和计算成本。Mistral Large 4 约 1 万亿总参数、490 亿激活参数,在 Mistral 位于欧洲的数据中心用 3800 块 NVIDIA Grace Blackwell GPU 训练,权重计划于 10 月底发布。

tools, enabling faster simulations and closed-loop AI-driven design. The partnership will introduce public benchmarks and open datasets to standardize evaluations for physics and agentic AI in computer-aided engineering, while leveraging NVIDIA’s Omniverse for 3D simulation and asset creation. Building on their existing work, including the development of Mistral Large 4, the initiative aims to push the boundaries of AI-powered engineering across industries like aerospace, automotive, and semico
匹配字段正文