MIT: GPT-6 edited a rival AI's weights by hand, lifting its math score to 48%

What's the deal? MIT CSAILDealroom has a profile for this one. Try Dealroom → researchers report that OpenAI's GPT-6 Astra directly edited the weights of Alibaba's small Qwen3.5-0.8B model — no retraining required — and raised its MATH-500 score from 7.9% to 48.1%. Astra changed just 0.016% of the model's weights, according to the paper published November 2, 2026.

How it works: Rather than running training loops or gradients, Astra read the raw weights and reasoned about what each part should compute. The researchers call this "Weight-Space Programming."

Why it matters: The edit targeted a flaw Qwen's own model card already flagged — repetitive reasoning loops. So this is better understood as a precise repair of a known failure than an AI independently learning new mathematics.

The upside: The fix generalised. The same weight correction improved instruction-following, tripling the model's IFEval score, the researchers say. Astra also identified which game a separate 17,253-weight network was trained on — Breakout — and wrote small game-playing networks from scratch.

What could go wrong? The result is preliminary. It comes from one small model, relies on repeated test-and-edit attempts, and has not been independently replicated or peer reviewed. Lower-budget attempts reportedly produced much smaller gains, leaving open whether the approach scales.

The signal: For years, neural network weights were shaped only by optimisers like Adam or by training code. The prospect here is different: capable language models that diagnose and patch other models by reasoning directly about their weights. If it holds up, it points to a new path for both optimisation and interpretability — though that remains a big if.

Read more: rikiyat.github.io