ModelsOct 4, 2026 · 16:106 sources

Aleph Alpha Kolibri Beats Qwen With 96.9% AIME Score

From huggingnews.com

Image: huggingnews.com
AI;DR

Aleph Alpha's Kolibri, a 78B-parameter open-weight model with a 1M token context window, scored 96.9% on the AIME 2025 math benchmark and 84.3% on GPQA Diamond, narrowly surpassing Qwen3.5 35B-A3B and GPT-OSS 120B in German evaluations, though it lagged behind Qwen3.6 in coding with 66.4% on SWE-Bench Verified versus 73.8%.

Read the sourceSign in to suggest

Sources

+ 4 more
  1. SUPPORT@testingcatalog“The Aleph Alpha team developed a bilingual German/English tokenizer and focused on including organic German data throughout the training process of the model, so that 21.3% of the pre-training tokens are German.”x.com
  2. SUPPORT@emollick“Europe's new “sovereign model” is fine-tuned on data generated by… GLM and Qwen”x.com
  3. SUPPORT@teortaxestex“Hybrid SWA, 3.46B active. Europe on the Pareto frontier indeed.”x.com
  4. SUPPORT@kimmonismus“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.”x.com
#llmMarkdown

One ranked edition, every morning

The stories that matter today. Free, cancel anytime.

Telegram @aidr_today