Live intelligence

Latest AI updates

1 updates
Models

Nous Research releases Arete-27B, an open Qwen3.8-27B fine-tune targeting political framing

Nous Research released Arete-27B, a supervised fine-tune of Qwen3.8-27B trained on 52.8 million tokens without reinforcement learning or architectural changes. On Nous Research's self-built 803-prompt evaluation, the reported rate of answers containing state-aligned framing fell from 22% to 0.9%, while the mean fact score increased from 2.32 to 2.98 out of 5. The weights, training data and evaluation materials are public, but the results have not been independently replicated and 216 evaluated answers still received a fact score of zero.

Arete-27BNous ResearchQwen3.8-27B