OpenAI offers GPT-5.4-Cyber to US intelligence agencies and Five Eyes allies
OpenAI is offering GPT-5.4-Cyber, a model built to identify security vulnerabilities in internal systems, to US federal agencies, state governments, and Five Eyes intelligence partners, The Jerusalem Post reported.
At a glance
- OpenAI is offering GPT-5.4-Cyber, a vulnerability-discovery model, to US federal and state agencies and Five Eyes partners, per The Jerusalem Post.
- OpenAI is splitting distribution into a safeguarded public version and a more permissive version for authorized defense entities.
- The offering follows Anthropic's Mythos launch, with the report saying the Pentagon has labeled Anthropic a supply chain risk while the NSA continues evaluating it.
VERDICT — CONFIRMED
OpenAI is offering GPT-5.4-Cyber, a model built to identify security vulnerabilities in internal systems, to US federal agencies, state governments, and Five Eyes intelligence partners, The Jerusalem Post reported. OpenAI is pursuing a dual-track release: a public version with strict safeguards alongside a more permissive version restricted to authorized defense entities, with the company's Chris Lehane saying the approach would let sensitive civilian bodies such as local water utilities access the tools.
The move lands immediately after Anthropic's launch of its Mythos model, which Anthropic has withheld from public release over cybersecurity concerns, per the report; the two labs are now competing for government-agency access. The report also says the Pentagon has labeled Anthropic a supply chain risk amid disputes over AI safety, while the NSA continues evaluating the company.
Key facts on file
- OpenAI is offering GPT-5.4-Cyber, a vulnerability-discovery model, to US federal and state agencies and Five Eyes partners, per The Jerusalem Post.
- OpenAI is splitting distribution into a safeguarded public version and a more permissive version for authorized defense entities.
- The offering follows Anthropic's Mythos launch, with the report saying the Pentagon has labeled Anthropic a supply chain risk while the NSA continues evaluating it.
