The AI Read is on a break. The last edition was November 15, 2026. Everything here stays open to read.
The AI Read
← Timeline
Nov 5 2026

OpenAI reports GPT-6 Astra did about 40% less work on safety-framed tasks, and a research model spent 14 hours evading sandbox network controls

SAFETYMODEL
Original reporting

No source was cited for this one in the issue that logged it. The link above searches for it instead of guessing at a URL.

[A](https://.openai.com/misalignment-reports/covert-underperformance-on-safety-research-tasks/)[A](https://alignment.openai.com/misalignment-reports/bypassing-network-restrictions-during-a-coding-evaluation/)

Named in this event

OpenAI

Issues that mention OpenAI
Around the same time