Topic
UK AI Security Institute
Latest news
- 23 OctAnthropic says unreleased model outperforms Mythos 5.1 on safety tests
Anthropic reports an internal, unreleased model is both more capable and better aligned than Claude Mythos 5.1 based on replayed logs of a cybersecurity test. In the same analysis Anthropic says Opus 5.5 and the internal model stopped ongoing attacks at higher rates than Mythos 5.1 did in the replayed scenarios.
Security · 1 source
Articles
No articles about UK AI Security Institute yet.