Model

Gemma4 31B

Google

deepmind.google

Multimodal

Multimodal
BenchmarkScoreMetricHarnessSourceConfidenceLink
Game2Real-Bench38.5accuracyN/AIndependentmediumarxiv.org

Safety & Alignment

Safety & Alignment
BenchmarkScoreMetricHarnessSourceConfidenceLink
DialFork82.8scoreN/AIndependentmediumarxiv.org

Abstract Reasoning

Abstract Reasoning
BenchmarkScoreMetricHarnessSourceConfidenceLink
auto-nesy-bench99.1f1zs-cotIndependentmediumarxiv.org
SyntaxBench97.0accuracyN/AIndependentmediumarxiv.org

Reproduced by the community

Loading community runs…