In a recent benchmark evaluation assessing performance against multi-stage cyberattacks, Anthropic's Claude Fable 5.2 outperformed GPT-6 Cyber. The comparison highlights the capabilities of advanced artificial intelligence models in complex cybersecurity scenarios, specifically testing their proficiency in handling multi-stage attack simulations. Further technical details regarding the benchmark metrics and specific test parameters were not detailed in the report.