OpenAI · Released 2026-11-05
A dedicated cybersecurity model: finding and validating exploitable vulnerabilities within researcher-defined scope, covering exploit development, vulnerability research, code review and patching. 48.5% on ExploitGym (from 32.1%), officially Critical for cybersecurity under the Preparedness Framework (up from High); access starts with approved Daybreak Red customers, expanding progressively. Not applicable to the AA / Pi / Physics-IQ boards - specialized cyber capability sits outside those calibers, and the approval-only gate leaves no independent third-party readings yet; the gap stays stated.
Original model explanation ↗No comparable leaderboard score is included yet. Explore the evaluations and examples below.
Official release and press analysis, not our test
OpenAI's dedicated cybersecurity model scores 48.5% on ExploitGym (from 32.1%), with the official Preparedness Framework moving cybersecurity up to Critical; the same model reproduced 1,724 Linux vulnerabilities and built 948 working exploits, and the 81 novel Firefox flaws are already shipping fixes via MFSA2026-118.
Official release · vendor evaluation
GPT-6's complex-work model connects understanding, tool use and file delivery.
Original release ↗Sol and Luna join the GPT-6 family with different capability and cost trade-offs.
Original release ↗Sol advances to 6.1, with reported gains in coding, computer use and professional work.
Original release ↗The same Luna model gains a dedicated endpoint for predicates, choices and scores; this is API access, not a separate set of model weights.
Original release ↗The same Sol model gains an API speed tier; its capability version and board identity remain, with separate cost and product-access conditions.
Original release ↗The formal guide read October 23 marks GA, still using gpt-6-luna. This records the observed documentation state, not a new model launch date. Input, refusals and billing connect to the Jev topic.
Original release ↗Terra advances to 6.1: professional computer tasks score +15 points over GPT-5.6 Terra at about a third of the cost, with standard API prices 50% lower; GitHub Copilot added it the same day. The AA board had no entry for it as of that day.
Original release ↗The behavior Spec is reorganized around five values: faithfulness to human intent, harm minimization, honesty and transparency, human agency and care. A new “consider context carefully” subsection, an updated chain of command, most detailed style guidance retired, and the model now called “GPT.” This is a Spec update, not a new model.
Original release ↗A dedicated cybersecurity model: 48.5% on ExploitGym (from 32.1%), Preparedness cybersecurity Critical (up from High); API and Codex first reach approved Daybreak Red customers, expanding progressively, with GPT-5.6 Cyber opening to all Daybreak Red at the same time. Outside the three boards; the gap stays noted.
Original release ↗