Model

Gemma-4-12B-it

Google

deepmind.google

Coding

Coding
BenchmarkScoreMetricHarnessSourceConfidenceLink
SolBehave57.9success_rateN/AIndependentmediumarxiv.org

Multimodal

Multimodal
BenchmarkScoreMetricHarnessSourceConfidenceLink
CREATE37.8accuracyN/AIndependentmediumarxiv.org
MVGraph20.5accuracyN/AIndependentmediumarxiv.org
Visual Sanad0.5discriminationN/AIndependentmediumarxiv.org

Safety & Alignment

Safety & Alignment
BenchmarkScoreMetricHarnessSourceConfidenceLink
LinguaGuardBench1.0flip_rateN/AIndependentmediumarxiv.org

Reproduced by the community

Loading community runs…