gpt-6POST
Public GPT-6 Astra said to refuse defensive cyber tasks less often
cramforce highlights that the public non-cyber GPT-6 Astra rarely refuses defensive cybersecurity tasks, suggesting it may be time to rerun deepsec.sh.
Malte Ubl
@cramforceArchived 2026.09.14Originally posted 2026.09.06
Malte Ubl @cramforceOriginally posted 2026.09.06
The important bit of news hidden in this benchmark release: GPT 6 Astra in the publicly available non-cyber version will perform defensive cyber security tasks without frequent refusals. It's the new best model, so it may be time to run https://deepsec.sh again.
This post quotes another post: x.com/vercel_dev/status/2096281878072635737
x.com/cramforce/status/2096283671586340967Archived 2026.09.14 16:12