OpenAI Enlists METR for Safety Review of Largest RL Run
November 16, 2026
1 min read
The evaluator will get model snapshots and internal records for a single-run pilot; a broader monitor assessment still awaits agreement on scope.... Read the full story at The Defiant