Code Generation on HumanEval (Accuracy, Peak Memory) benchmark leaderboard · SOTA2 Research

Code Generation on HumanEval (Accuracy, Peak Memory)

34.2AccuracyLDAdam

Evaluation Results

Method	Links
LDAdam2026.04		34.2	69.37
Adam2026.04		34.01	93.26
OASIS2026.04		33.5	46.35
LoRA2026.04		31.84	62.08
GaLore2026.04		31.24	69.2
OASIS2026.04		29.28	46.16
LoRA2026.04		28.34	61.7
LDAdam2026.04		27.93	68.99
GaLore2026.04		27.42	69