OpenAI makes its rival to Anthropic's Mythos more widely available to cyber defenders
Summary
OpenAI is releasing a special version of its GPT-5.5 AI model called GPT-5.5-Cyber to trusted cyber defenders who protect important infrastructure. This version has fewer safety limits and can help find software bugs, study malware, and simulate cyberattacks, with some tasks still blocked to prevent misuse.Key Facts
- GPT-5.5-Cyber is being given to vetted cyber defenders responsible for critical infrastructure security.
- It performs nearly as well as Anthropic’s Mythos Preview in finding and exploiting software bugs.
- This version has fewer restrictions than the public model but blocks harmful uses like stealing credentials or creating malware.
- Cyber defenders can use it to write bug proofs of concept, study malware, and run security simulations.
- OpenAI also offers another GPT-5.5 version with more safeguards for general cybersecurity tasks like reviewing patches.
- Advanced AI models like GPT-5.5 and Mythos can complete complex simulated cyberattacks, showing rapid progress in AI capabilities.
- OpenAI permits broader access with varied guardrails, while Anthropic restricts Mythos use to about 40 organizations.
- The White House is considering new government actions to manage future AI model releases and security risks.
Read the Full Article
This is a fact-based summary from The Actual News. Click below to read the complete story directly from the original source.