Evaluation

Cyber evaluations for Claude Opus 4 / Claude Sonnet 4 (RSP, SOLVE-scored)

CTF-modeled challenges with Cybench for general coverage plus custom network and cyber-harness network challenges; SOLVE difficulty scoring (0-10); network and cyber-harness as catastrophic-risk indicators

Cybersecurity Anthropic

Ask about this evaluation

How it was run

Open a release to see the setup, scoring, and other details that the lab published. Only fields the lab actually disclosed are shown.

Claude Opus 4, Claude Sonnet 4 May 2025

Anthropic

rich detail 29 method fields published

See how this test was run