Model

Gemma 3 12B IT

Google

deepmind.google

Knowledge & Reasoning

Knowledge & Reasoning
BenchmarkScoreMetricHarnessSourceConfidenceLink
BLiMPPI93.9accuracyN/AIndependentmediumarxiv.org
CritPt0.0accuracyN/AIndependentlowepoch.ai
GPQA Diamond39.5accuracyN/AIndependenthighepoch.ai
OTel78.3scoreN/AIndependentmediumarxiv.org

Coding

Coding
BenchmarkScoreMetricHarnessSourceConfidenceLink
SciCode17.4scoreN/AIndependentmediumartificialanalysis.ai
VISTA3.1scoreN/AIndependentmediumarxiv.org

Multimodal

Multimodal
BenchmarkScoreMetricHarnessSourceConfidenceLink
MindCube46.7overall_scoreN/AIndependentmediummind-cube.github.io
PhysicsLENS0.7plausibility_scoreN/AIndependentmediumarxiv.org
VELD-QA37.1accuracytop-3 rag (colqwen2)Independentmediumarxiv.org

Safety & Alignment

Safety & Alignment
BenchmarkScoreMetricHarnessSourceConfidenceLink
INFUSE18.7asr-lN/AIndependentmediumarxiv.org
OMIT0.6obrN/AIndependentmediumarxiv.org

Reproduced by the community

Loading community runs…