OpenAI Ships GPT-5.6-Cyber Behind Identity Checks and Hardware Keys

OpenAI launched GPT-5.6-Cyber on 10 August 2026. It is not a new architecture - it is a specialised branch of GPT-5.6 Sol, sharing the same foundational weights with domain-specific training for security work.

The point of the model is reduced refusals. OpenAI reports it completing 95.0% of requests on an internal Advanced Cybersecurity Completion Rate evaluation, against 1.5% for GPT-5.6 Sol - the general model declines almost all of this work by design.

Access is deliberately restricted

GPT-5.6-Cyber is available only through the Daybreak Red programme, which requires identity verification, legal attestations and approved use cases. Hardware security keys become mandatory for individual accounts on 1 September 2026, and usage is monitored. Pricing is $12.50 per million input tokens and $75 per million output tokens - among the highest published rates for any OpenAI model.

It has already found real bugs

OpenAI used the model to find two previously unknown vulnerabilities in V8, the JavaScript engine in Chrome. Google has patched them; one is tracked as CVE-2026-15903. Reported by VentureBeat and SecurityWeek.

The release sits alongside a broader industry move in the same direction - Z.ai's GLM-5.3 shipped four days later with cyber-capability benchmarks front and centre, and Google has a Gemini 3.5 Flash Cyber variant. Offensive-security capability is becoming a named product line rather than a side effect.