Aleph Alpha发布Kolibri：78B参数高效模型，德语英语表现强劲
Aleph Alpha’s Kolibri looks like a surprisingly strong AI release from Germany!…
非中美巨头发布的欧洲原生强效 MoE 模型，参数效率极高且开源，对关注多语言及轻量化部署的团队有参考价值。
Aleph Alpha’s Kolibri looks like a surprisingly strong AI release from Germany! I did not expect that but respect where respect is due!
Aleph Alpha 的 Kolibri 看起来像是来自德国的一款实力惊人的 AI 发布！我没想到这一点，但值得尊重！
78B total parameters, just 3.46B active per token, and open weights under Apache 2.0. Built from scratch for German and English.
总参数 78B，每个 token 仅激活 3.46B，以 Apache 2.0 许可证开源权重。专为德语和英语从头构建。
In Aleph Alpha’s own evaluations, it scores 96.9% on AIME 2025 and 84.3% on GPQA Diamond. Its German overall score narrowly beats Qwen3.5 35B-A3B and GPT-OSS 120B.
在 Aleph Alpha 自身的评估中，它在 AIME 2025 上得分 96.9%，在 GPQA Diamond 上得分 84.3%。其德语综合得分 narrowly beats Qwen3.5 35B-A3B 和 GPT-OSS 120B。
Coding is more mixed: 66.4% on SWE-Bench Verified, behind Qwen3.6’s 73.8%.
编程方面表现不一：在 SWE-Bench Verified 上得分为 66.4%，落后于 Qwen3.6 的 73.8%。
It increasingly looks as though any "moat" - especially with small models - is evaporating, and it no longer matters whether the players are from China, the US, or Europe. Anyone can train good small models.
现在看来，任何“护城河”——尤其是对于小模型而言——正在逐渐消失，玩家是来自中国、美国还是欧洲已不再重要。任何人都可以训练出色的小模型。
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