Development

A cyber model completes 95% of advanced tasks against 1.5% for its civilian sibling

development2026-08-12

Claim

OpenAI split its trusted-access program into tiers and unveiled GPT-5.6-Cyber, completing 95% of advanced cyber tasks against 1.5% for its civilian sibling, and already finding two chained zero-days in a browser engine.

Reported metric: 95% vs 1.5%

Reporting date: 2026-08-12 (not necessarily the event date)

Sources

Attribution and reporting linked by this record; links do not imply independent verification.

Welcome to August 12, 2026 (theinnermostloop.substack.com)

Knowledge graphopen in browser →

Reported in

Welcome to August 12, 2026

Actor

Evidence for

Supersedes

A lab slows a release because it cannot rule out Critical cyber capability

Referenced by

Cite or correct this record

Citation formatCitation text

A cyber model completes 95% of advanced tasks against 1.5% for its civilian sibling
Innermost Loop KB — secondary record
Reporting date: 2026-08-12
KB permalink: https://nicholsn.github.io/innermost-loop-kb/developments/2026-08-12-a-cyber-model-at-ninety-five-percent-versus-one-point-five
Claim: OpenAI split its trusted-access program into tiers and unveiled GPT-5.6-Cyber, completing 95% of advanced cyber tasks against 1.5% for its civilian sibling, and already finding two chained zero-days in a browser engine.
Attributed sources:
- Welcome to August 12, 2026 — https://theinnermostloop.substack.com/p/welcome-to-august-12-2026
This citation does not imply independent verification of linked-source claims.

Suggest a correction — opens a GitHub draft for you to review and submit.