DeepSeek V3 on MMLU (Massive Multitask Language Understanding): Score Breakdown

DeepSeek V3 × MMLU (Massive Multitask Language Understanding)

Rincian Skor

Peringkat ModelSkor
#5 / 20588.5

Analisis Kinerja

DeepSeek V3 achieves 88.5 on MMLU (Massive Multitask Language Understanding).