Functional Correctness on HumanEval (p@k and Precision) benchmark leaderboard · SOTA2 Research

Functional Correctness on HumanEval (p@k and Precision)

71.57Pass@1Dense

Evaluation Results

Method	Links
Dense2026.06		71.57	85.58	89.63	100
SecRL-Prune2026.06		54.79	73.18	84.21	85.3
PruneNet2026.06		46.43	62.08	70.34	72
SecRL-Prune2026.06		41.95	60.78	78.65	72.5
SecRL-Prune2026.06		38.16	57.38	75.14	68.1
Dense2026.06		36.7	68.3	87.8	100
SecRL-Prune2026.06		34.37	53.98	71.63	63.7
Dense2026.06		27.67	60.04	83.53	100
SecRL-Prune2026.06		26.2	53.5	78.04	79.5
SecRL-Prune2026.06		21.37	46.91	72.25	80.6
PruneNet2026.06		21.3	38.57	51.05	43.9
SecRL-Prune2026.06		18.94	35.93	58.39	56.9
PruneNet2026.06		18.7	34.71	47.29	39.8
PruneNet2026.06		16.1	30.85	43.53	35.7
SecRL-Prune2026.06		15.32	31.58	53.68	49.7
SecRL-Prune2026.06		15.06	33.78	60.97	61.2
PruneNet2026.06		14.96	33.07	49.85	56.3
PruneNet2026.06		14.16	30.09	42.4	43.6
SecRL-Prune2026.06		11.91	27.72	55.33	51.8
SecRL-Prune2026.06		11.7	27.23	48.97	42.5
PruneNet2026.06		10.55	20.77	33.1	32.3
SecRL-Prune2026.06		8.75	20.65	49.69	41.8
PruneNet2026.06		7.98	20.5	37.19	35.8
PruneNet2026.06		7.9	15.04	23.8	23.6
PruneNet2026.06		5.96	14.73	29.88	27.3
PruneNet2026.06		5.8	10	17.6	16.8
PruneNet2026.06		4.2	9.7	23.85	20