# MIT researchers show AI can alter model weights without retraining

- URL: https://talkfuse.com/story/forget-retraining-ai-can-now-tweak-another-ais-weights/
- Published: 2026-11-03T08:15:00+00:00
- Updated: 2026-11-03T08:32:20+00:00
- Original publisher: The Register
- Original article: https://www.theregister.com/ai-and-ml/2026/11/03/forget-retraining-ai-can-now-tweak-another-ais-weights/5307105
- Topics: AI, Space, Technology

## Key points

- GPT-6 Astra directly edited another model’s weights.
- A small change boosted maths benchmark performance up to sixfold.
- Researchers say standard testing is the next step.

## Summary

MIT CSAIL researchers report that a frontier AI model, GPT-6 Astra, can directly edit another model’s numerical weights to change its behaviour, without retraining. They call the approach “Weight-Space Programming”, and say it could offer new ways to improve and understand models.

In a demonstration, Astra changed 0.016 per cent of Qwen3.5-0.8B’s weights, improving its score on the MATH-500 benchmark by up to sixfold and roughly tripling its IFEval instruction-following score. The researchers also say Astra wrote weights from scratch for a Breakout-playing policy network. They see possible applications in reducing catastrophic forgetting and improving interpretability and alignment, but say the next step is to develop a standard benchmark and testing framework.

---
Summarised by TalkFuse (https://talkfuse.com/). Quote the original publisher for the reporting and link this page for the summary and discussion.