OpenAI is gating GPT-5.5 Cyber to credentialed defenders first, mirroring the Anthropic Mythos rollout Altman publicly criticized.
Key Takeaways
GPT-5.5 Cyber supports pen testing, vulnerability identification and exploitation, and malware reverse engineering.
Access requires an application with credentials and planned-use details; broader rollout contingent on U.S. government consultation.
Altman called Anthropic’s identical Mythos gating “fear-based marketing” before replicating the same policy days later.
An unauthorized group reportedly bypassed Mythos restrictions anyway, undermining the security rationale for both approaches.
Hacker News Comment Review
Commenters read the Altman reversal as a pure positioning play, comparing it to Goldman Sachs changing its tune after taking a position.
Skepticism about Altman’s credibility dominated the thread; the pattern is seen as habitual, not an isolated contradiction.
One commenter raised a separate concern: subsidy removal could spike token prices, making these tools economically inaccessible regardless of credential gating.
Notable Comments
@feverzsj: flags that expiring subsidies may cause token prices to spike, a cost barrier independent of access policy.
