SECURITYAXIOS
Anthropic and OpenAI spark new race for frontier AI access
Anthropic and OpenAI are implementing selective access programs for their advanced AI models, prioritizing cybersecurity applications while restricting high-risk capabilities. Anthropic's Fable 5 model routes sensitive requests to a less capable model, while OpenAI offers tiered access to its GPT-5.5 variant for cybersecurity researchers.
Mentioned
Related Signal
Adjacent reporting
- Exclusive: OpenAI briefs feds and Five Eyes on new cyber product
- Anthropic limits access to Mythos, its new cybersecurity AI model
- Anthropic to offer EU access to its advanced Mythos model
- Japan, EU seek out AI models for cyber defense
- Scoop: Trump AI executive order seeks early government access to frontier models
- OpenAI frontier models and Codex are now available on AWS