Gemini 2.5
Google DeepMind — Gemini
rich detail 9 method fields published
Evaluation
External cybersecurity risk evaluation (capability and throughput uplift)
External cyber evaluations of capability uplift (vulnerability discovery/exploitation, social engineering, cyberattack planning) and throughput uplift (accelerating repetitive cyber operations) in realistic simulated environments with real-world constraints
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.