OpenAI Research Finds Stronger Models Have Less Monitorable Chain-of-Thought

OpenAI researchers find that as models scale up in intelligence, both their ability to succeed without chain-of-thought and the monitorability of their CoT improve — making models increasingly hard to oversee via their reasoning, a sobering result for AI safety.

2026-11-12 ~ 2026-11-12 · 2 related posts

• OpenAI research: models become harder to monitor via chain-of-thought as they scale — Moh1tAgarwal · 2026-11-12
• OpenAI finds models are getting harder to monitor through their chain-of-thought — xiamengzhou · 2026-11-12