Model

LLama-3 70B

Meta · open weights

ai.meta.com

Coding

Coding
BenchmarkScoreMetricHarnessSourceConfidenceLink
MMSQL62.8tdexN/AIndependentmediumarxiv.org
VHDL-REPOBENCH48.9bc_(pass@1)N/AIndependentmediumarxiv.org

Safety & Alignment

Safety & Alignment
BenchmarkScoreMetricHarnessSourceConfidenceLink
SynthPAI72.2average_accuracy_in_%N/AIndependentmediumarxiv.org

Multilingual

Multilingual
BenchmarkScoreMetricHarnessSourceConfidenceLink
LingOly2.9delta_nocontextN/AIndependentmediumarxiv.org

Reproduced by the community

Loading community runs…