GPT-5.6 Cyber

OpenAI's first purpose-trained cybersecurity SKU, released August 10, 2026 as `gpt-5.6-cyber` (alias `daybreak-red-latest`) when the Daybreak program split into two tiers: Daybreak Blue (frontier general models with defence-calibrated safeguards, alias pointing at GPT-5.6 Sol) and Daybreak Red (cyber specialists for authorized vulnerability research, exploit validation and security testing). Access is the product: approval, identity verification, legal attestation, monitoring, and — from September 1, 2026 — mandatory hardware security keys on individual accounts. $12.50/$1.25 cached/$75 per Mtok against Sol's $5/$0.50/$30, with a smaller 400K context and Chat Completions unsupported. Rated **High**, not Critical, on OpenAI's cyber preparedness scale (Astra later became the first Critical model). NO BENCHMARK ROWS ARE RECORDED: OpenAI published no numeric public benchmark table for this SKU, only directional claims (beats Sol on ExploitGym and internal zero-day work; LAGS Sol on vulnerability discovery and report writing; Sol is more token-efficient on ExploitBench 300-turn), plus a press-relayed internal figure of ~95% dual-use request completion vs ~1.5% for safeguarded Sol. Sol's general leaderboard scores are not this model's scores, and no independent evaluation exists yet — so the entry ships spec-only and stays unranked rather than borrowing numbers.