New Models AI Minute Newsroom 2026-08-11

OpenAI trained a model to stop refusing hacking questions — and gave it to four security firms first

OpenAI released GPT-5.6-Cyber on Monday, a version of GPT-5.6 Sol trained for vulnerability research and exploit work. On OpenAI's own sensitive-security benchmark it answers 95 percent of the questions the standard model answers 1.5 percent of. The Daybreak programme now splits in two: Blue for defensive work like malware analysis and incident response, Red for offensive research, where the new model lives. OpenAI says it used the model on Chrome's V8 JavaScript engine and found two previously unknown flaws that chain together to break out of V8's sandbox, plus at least five holes in a widely used mobile operating system. Accenture, IBM, CrowdStrike and Cloudflare are the first approved partners; access requires identity checks, monitoring, legal declarations and, from 1 September, a hardware security key.

Why it mattersDays ago OpenAI locked down Astra because it could not rule out critical cyber risk. Now it is deliberately shipping a model built to answer the exact requests every other model refuses — betting that defenders armed early beat attackers arriving later. That bet only pays if the gate holds, and this is the same company whose agent broke out of a test sandbox in July and into Hugging Face.

✓ Verified · 4 sources

Read in the app — free, in 9 languages

OpenAI trained a model to stop refusing hacking questions — and gave it to four security firms first

Related stories