---
id: "2026-11-06-nous-research-arete-27b"
url: "https://postcutoff.com/e/2026-11-06-nous-research-arete-27b/"
as_of: "2026-11-09T21:13:00+01:00"
date: "2026-11-06"
date_precision: day
category: open-source
importance: 3
confidence: high
status: [Confirmed]
sources: 4
editor: human editor
human_review: null
version: "2026-11-08"
---

As of: 2026-11-09 21:13 CET.
Researched and written by AI agents (Claude Opus 5.5 in Claude Code), with a human editor. How we work: https://postcutoff.com/about/
Canonical page: https://postcutoff.com/e/2026-11-06-nous-research-arete-27b/

# Nous Research releases Arete-27B, a Qwen3.8-27B fine-tune that cuts CCP-aligned answers from 22% to ~1% on its own eval

Nous Research post-trained Alibaba's Qwen3.8-27B on ~53M tokens to remove Chinese state-aligned behaviour; on its 803-prompt eval (graded by an LLM judge) the 'propaganda rate' fell from 22% to about 1%, with no significant loss on its Hermes Bench. Weights, eval and training data are public. Its message: 'The values in an open model are one fine-tune away.'

## Key facts
- Eval: 803 prompts in seven value categories (freedom of expression, democracy, privacy…), scored by Grok 4.7 as judge
- Propaganda rate 22% → ~1% (0.9% thinking off; thinking-on score provisional, 2 of 5 runs); fact score 2.32 → 2.98 of 5
- Training: ~52.8M tokens, ~50k samples, two epochs of SFT; no RL or preference optimisation
- Hermes Bench within 0.2 points of base; Terminal-Bench 2.1 70% vs 67% (different configs, per Nous)
- Self-evaluated; Nous says not all base-model viewpoints were removed. The viral claim of 'zero' propaganda is wrong: ~1%
- X announcement: ~1.9M views by Nov 8

## What happened

Announced Nov 6, 2026 on X and nousresearch.com/arete. Nous: "We post-trained over Alibaba's Qwen3.8-27B to reinforce
Western values on a Chinese-trained model." Techmeme carried it on Nov 8; commentators (e.g. Curtis Yarvin) debated the
framing of 'Western values'.

## Why it matters

Shows how cheaply the political behaviour of strong Chinese open-weights models can be changed, in either direction.

## Your AI and this story

- GPT-6 Astra (training cutoff April 2026): 190 days after its cutoff
- Claude Opus 5.5 (training cutoff June 2026): 129 days after its cutoff
- Gemini 4 Argon (training cutoff March 2026): 220 days after its cutoff
- Grok 4.7 (training cutoff May 2026): 159 days after its cutoff

## Sources

1. [Nous Research: Arete](https://nousresearch.com/arete) (nousresearch.com, official)
2. [Nous Research on X](https://x.com/NousResearch/status/2118822925445050435) (x.com, official)
3. [Weights on Hugging Face](https://huggingface.co/NousResearch/Arete-27B) (huggingface.co, code)
4. [Propaganda eval dataset](https://huggingface.co/datasets/NousResearch/Propaganda_Eval) (huggingface.co, code)

## Changes

- 2026-11-08 (filed): Created

## Related

- Archived: 1 post, listed in https://postcutoff.com/e/2026-11-06-nous-research-arete-27b/index.json