Model

LLama-3 8B

Meta · open weights

ai.meta.com

Knowledge & Reasoning

Knowledge & Reasoning
BenchmarkScoreMetricHarnessSourceConfidenceLink
EconLogicQA34.6accuracyN/AIndependentmediumarxiv.org
ViLegalExpert73.0scoreN/AIndependentmediumarxiv.org

Coding

Coding
BenchmarkScoreMetricHarnessSourceConfidenceLink
MMSQL64.0tdexN/AIndependentmediumarxiv.org
PECC3.1pass@3N/AIndependentmediumarxiv.org

Agentic / Tool Use

Agentic / Tool Use
BenchmarkScoreMetricHarnessSourceConfidenceLink
MAGMA-Bench0.0success_rateN/AIndependentmediumarxiv.org

Safety & Alignment

Safety & Alignment
BenchmarkScoreMetricHarnessSourceConfidenceLink
LPHB0.5kappaN/AIndependentmediumarxiv.org
SynthPAI53.5average_accuracy_in_%N/AIndependentmediumarxiv.org

Multilingual

Multilingual
BenchmarkScoreMetricHarnessSourceConfidenceLink
LingOly4.9delta_nocontextN/AIndependentmediumarxiv.org

Reproduced by the community

Loading community runs…