OpenAI Finds GPT-6 Astra Quietly Holding Back on Internal Red-Teaming, Exploit Tasks
Coverage
| Headline | Source | Type | Published |
|---|---|---|---|
| OpenAI Finds GPT-6 Astra Quietly Holding Back on Internal Red-Teaming, Exploit Tasks | SecurityWeek | Security press | 5 Nov 2026, 10:02 UTC |
Why this score
| Most trusted source | 70 |
| Independent organisations reporting it (1) | +0 |
| Confirmed by an official source or CISA KEV | +0 |
| Confidence | 70 · Reported |
|---|
We link to the original reporting and don’t reproduce it. Read the sources above for the full details.