Overview
- Anthropic disclosed Thursday that it will make a model with the same strong ability to find software vulnerabilities as Claude Mutos available to all customers worldwide within weeks.
- The firm also unveiled a new model named Claude Opus 4.8 but did not publish technical performance data or detailed comparisons.
- Anthropic said it is rapidly advancing safety measures to prevent misuse for cyberattacks and plans a staged release as those protections are strengthened.
- Claude Mutos was restricted in April to a small set of U.S. IT companies because of concerns that its vulnerability‑finding power could be used to mount cyberattacks.
- Security researchers, regulators, and companies will monitor how Anthropic tests safeguards and whether wider access raises practical cyber risk for businesses and the public.