Nous Research Releases Arete-27B to Reshape Open-Model Values

Nous Research has released Arete-27B, an open-source post-trained version of Qwen3-27B designed to shift the value alignment of open-weight models, reportedly reducing Chinese-model propaganda rates from 22% to 1% across a seven-category value benchmark.

2026-11-07 ~ 2026-11-07 · 2 related posts

• Nous Research releases Arete-27B, cutting a Chinese open model's propaganda rate from 22% to 1% — NousResearch · 2026-11-07
• Nous Research ships Arete-27B, a Qwen3-27B fine-tune trained to reinforce Western values — altryne · 2026-11-07