Model Releases

We recently found some instances of CoT grading during the training of previously deployed models after building a system that scans all Ope…

We recently found some instances of CoT grading during the training of previously deployed models after building a system that scans all OpenAI RL runs for accidental CoT grading. We did not find clea

DGX agentx-post
model-releasesopenai--x

We recently found some instances of CoT grading during the training of previously deployed models after building a system that scans all OpenAI RL runs for accidental CoT grading. We did not find clear evidence that these instances degraded CoT monitorability.

Source: OpenAI (X) | 2026-05-07

Loading related sources…