Mistral Large 4 发布：1T参数欧洲训练
This is a surprising great release. Le Chaton fat is real! Mistral Large 4 puts…
Mistral Large 4 是全新旗舰模型发布，关键能力指标直接对标当前主流模型，且训练硬件配置极具讨论价值，值得开发者关注。
This is a surprising great release. Le Chaton fat is real! Mistral Large 4 puts Europe back in contention on coding and cybersecurity.
这是一次令人惊喜的发布。Le Chaton 模型实力强劲！Mistral Large 4 让欧洲在编码和网络安全领域重新具备竞争力。
Didnt expect Mistral to compete with GLM5.3!
没想到 Mistral 能与 GLM5.3 竞争！
“Le Chonk” is a natively multimodal model with 1T parameters and 49B active. The preview already delivers:
“Le Chonk”是一个原生多模态模型，拥有 1T 参数和 49B 激活参数。预览版已展现出以下能力：
- DeepSWE v1.1: 61.7%, roughly level with GLM-5.3 at 61%. Kimi K3 remains ahead at around 68% in Mistral’s comparison. - Artificial Analysis Cyber Index: 50, matching GLM-5.3-Flash and beating GLM-5.3’s 36. - CyberGym-E2E-AA: 82%, ahead of MiMo-V2.6-Pro’s 79%.
- DeepSWE v1.1：61.7%，与 GLM-5.3 的 61% 大致持平。在 Mistral 的对比中，Kimi K3 仍以约 68% 领先。 - Artificial Analysis 网络安全指数：50，与 GLM-5.3-Flash 持平，并优于 GLM-5.3 的 36。 - CyberGym-E2E-AA：82%，高于 MiMo-V2.6-Pro 的 79%。
Mistral says it trained the model from scratch in its own European datacenters. NVIDIA says only 4,000 Grace Blackwell Superchips powered the training! Which is impressive.
Mistral 表示该模型是在其自有欧洲数据中心从头训练的。NVIDIA 称训练仅使用了 4,000 颗 Grace Blackwell Superchip！这令人印象深刻。
Learning: you can compete with much less compute. Which is crazy. Congrats Mistral!
启示：你可以用更少的算力实现竞争。这太疯狂了。恭喜 Mistral！
更进一步：量化金融体系
看懂新闻只是起点——沿量化金融路径，把它变成能交付的工程能力