MIT researchers report GPT-6 Astra editing a small model's weights directly

Rikiya Takehi, Jayden Teoh, Alex Zhang and Omar Khattab of MIT CSAIL report that GPT-6 Astra, reading and editing weights without gradient descent, changed 0.016 percent of Qwen3.5-0.8B and raised its MATH-500 score from 7.9 to 48.1 percent. The authors call the results early and say the edit removed an obstacle rather than adding ability.