Google reports 57.0% for Gemini 4 Hydrogen on Terminal-Bench 4.1, a cheating-adjusted average reward from its own runs, against 18.2% for Gemini 3.8 Flash.
Tasks: Cheating-adjusted accuracy, as Google labels it.
Setting: High effort level, per Google's methodology; no numeric budget is stated.
Limits: Provider-reported: Google publishes the result for its own model; it has not been independently reproduced here. Google says the scores are as of November 2026 for the model it calls Gemini 4 Hydrogen; equivalence to the build now served on every surface is not established. Results under the same benchmark name can be shown together as reported scores, but unmatched settings and missing methods rule out a controlled ranking.
Google DeepMind · Google DeepMind · Google DeepMind · Google DeepMind · Google · Google
