德国 Aleph Alpha 发布开源小模型 Kolibri：78B 总参数、3.46B 激活
德国人自己从零训了个开源小模型，德语成绩还压过 Qwen3.5 和 GPT-OSS，数学也够硬，编程稍弱。
德国公司 Aleph Alpha 发布开源模型 Kolibri，总参数 78B，每 token 仅激活 3.46B，权重以 Apache 2.0 协议开放，从零针对德语和英语训练。官方评测显示其在 AIME 2025 得分 96.9%，GPQA Diamond 得分 84.3%，德语综合成绩小幅领先 Qwen3.5 35B-A3B 和 GPT-OSS 120B。编程方面表现一般，SWE-Bench Verified 得分 66.4%，落后于 Qwen3.6 的 73.8%。
Aleph Alpha’s Kolibri looks like a surprisingly strong AI release from Germany! I did not expect that but respect where respect is due!
78B total parameters, just 3.46B active per token, and open weights under Apache 2.0. Built from scratch for German and English.
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.
Coding is more mixed: 66.4% on SWE-Bench Verified, behind Qwen3.6’s 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.
- rohanpaul_ai10-05 22:29原文
