Evaluation

External evaluations of Gemini Ultra (autonomous replication, CBRN, cyber, societal)

Structured evaluations, qualitative probing and unstructured red teaming by independent external groups on a December 2023 Gemini API Ultra checkpoint, covering autonomous replication, CBRN, cyber, and societal risks (representational/distributional harms, neutrality/factuality, robustness/information hazards)

External & third-party testing Google DeepMind — Gemini

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.

Gemini 1.0 Dec 2023

Google DeepMind — Gemini

rich detail 13 method fields published

See how this test was run