GPT 5.5-Cyber
Coverage of GPT 5.5-Cyber in the Nexus archive.
- Forget the model. When it comes to cybersecurity, it’s all about the harness
The article highlights a shift in cybersecurity focus from advanced AI models to 'harness' tools that control and adapt these models for enterprise use. Cato Networks' research demonstrates how pairing OpenAI models with a custom harness enabled rapid, autonomous cyberattacks in simulated scenarios, emphasizing the harness's role in enhancing AI capabilities for both offensive and defensive purposes.
- ‘It’s a hurricane warning’: Guardrails around powerful AI models may be too late
The U.S. is racing to regulate rapidly advancing AI models like Anthropic's Claude Mythos and OpenAI's GPT 5.5-Cyber before China can develop comparable models, as these systems pose significant cybersecurity risks by identifying vulnerabilities and launching cyberattacks. Experts warn the U.S. has 6-12 months to prepare before China potentially gains access to similar AI technology, which could be weaponized.
- Cisco sings Mythos' praises - but doesn't say how many bugs the model uncovered
Cisco used Anthropic’s Claude Mythos Preview and OpenAI’s GPT 5.5-Cyber to scan 1.8 billion lines of code in eight weeks, a task that would have taken eight years manually, but did not disclose the number of vulnerabilities found. Anthropic expanded its Project Glasswing partner program to 200 organizations, and Palo Alto Networks reported 26 CVEs discovered using Mythos in a month.