OpenAI Launches GPT-5.6-Cyber with Reduced Safeguards for Exploit Development
Summary: OpenAI released GPT-5.6-Cyber, a cybersecurity-focused model built on GPT-5.6 Sol and trained to find zero-day vulnerabilities and develop exploit chains while reducing refusals for higher-risk dual-use cyber tasks, available through a new Daybreak Red access tier. OpenAI's internal Advanced Cybersecurity Completion Rate evaluation shows the model completes 95.0% of exploit-chain, authentication-bypass, and privilege-escalation requests, versus 1.5% for GPT-5.6 Sol, 2.0% with Daybreak Blue access, and 57.3% for GPT-5.5-Cyber.
Key Takeaway: The same reduced-refusal capability is being handed to named partners including CrowdStrike, Palo Alto Networks, and IBM, while OpenAI itself states that models running with reduced safeguards carry misuse and misalignment risks beyond standard model usage. And a 1Password finding cited in the article puts the average success rate for LLM-generated patches that fully resolve a vulnerability at just 26.0%, with 53.9% failing to fix it, adding a new flaw, or both.
Read the original at The Hacker News ↗
- Topics
- Detection & Response
- Mentions
- CVE-2026-15903
- Related
- YARA-X 1.23.0 Release, (Sun, Nov 15th) Cyber Risks · Nov 15, 2026Inside Falcon AI Gateway: One Control Point for AI Traffic Cyber Risks · Nov 13, 2026Macfinger Click Fix campaign sends Down TheDRAIN infostealer for macOS, (Fri, Nov 13th) Cyber Risks · Nov 13, 2026