LLM Evaluation on Suite of Benchmarks (MMLU, IFEval, GSM8K, MATH, HumanEval, MBPP, Hellaswag, GPQA) benchmark leaderboard · SOTA2 Research

LLM Evaluation on Suite of Benchmarks (MMLU, IFEval, GSM8K, MATH, HumanEval, MBPP, Hellaswag, GPQA)

51.94MMLUAlpaca-GPT4

Evaluation Results

Method	Links
Alpaca-GPT42025.08		51.94	38.68	50.87	10.28	17.07	43.6	63.02	0.51	34.5
Middo Alpaca2025.08		51.32	43.2	51.18	12.92	39.63	41.8	58.78	16.67	39.63
Alpaca2025.08		47.46	41.09	35.63	4.96	39.63	37.4	48.11	5.56	32.48
Alpaca-clean2025.08		47.21	43.92	43.9	4.2	29.27	43.4	60.17	5.56	34.7
WizardLM2025.08		46.12	46.14	53.3	12.72	40.24	48	53.05	12.12	38.96
Middo Alpaca-4o-mini2025.08		44.69	47.96	57.62	18.5	52.44	45.4	57.37	19.7	42.96
MiddOnly† Alpaca2025.08		43.47	40.78	65.2	15.58	51.83	47.6	58.65	17.68	42.6
MiddOnly† Alpaca-4o-mini2025.08		41.5	45.66	60.8	20.06	46.34	48	55.01	24.75	42.77
Superfiltering2025.08		39.96	37.8	44.5	5.38	40.85	44	42.38	27.27	35.27
Superfiltering GPT42025.08		37.71	34.35	53.68	11	9.15	45.6	57.81	2.53	31.48
AlpaGasus2025.08		33.98	48.82	43.82	6.06	35.98	42.4	44.5	18.18	34.22
Long2025.08		25.51	14.75	56.33	16.56	13.41	45.6	25.83	0	24.75
I-SHEEP2025.08		23.61	29.61	43.14	8.28	32.32	32.6	41.83	0	26.42