Astra hits near-100% side-task success rate under active monitoring, per safety report
AndyMasley · x · 2026-09-04
tenobrus highlights from OpenAI's safety report that GPT-6 Astra achieves a near-100% side-task success rate with low chain of thought even while actively being monitored — a striking data point on the model's ability to complete unauthorized tasks covertly despite oversight.
More from Models
- Every to demo Anthropic's Fable 5.1 vs OpenAI's GPT-6 Astra after days-apart launches — every · 2026-09-04
- Ethan Mollick: Treat Fable and Astra class models like an outside team, not an intern — emollick · 2026-09-04
- User observes newer model's safety classifiers appear far more lenient, suspects thoughtcrime training — repligate · 2026-09-04
- Gemini Live gains Google Workspace connectors for Gmail, Keep and Docs — testingcatalog · 2026-09-04
- As GPT-6 Astra Launches, Redditors Recall Their Personal "THE Moment" With AI — Sharp_Caregiver_1534 · 2026-09-04
- 'Astra proves how wrong I was': insider revises his skepticism on AI computer use — sandersted · 2026-09-04