Commonsense Reasoning and Question Answering on ARC-e, ARC-c, Winogrande, BoolQ, PIQA, HellaSwag, OBQA, HQA benchmark leaderboard · SOTA2 Research

Commonsense Reasoning and Question Answering on ARC-e, ARC-c, Winogrande, BoolQ, PIQA, HellaSwag, OBQA, HQA

60.5ARC-e AccPythia

Evaluation Results

Method	Links
Pythia2026.04		60.5	31.2	61.3	61.1	71.1	53.6	33.2	31.9	50.49	1,237.4	—	632.48
OPT2026.04		57.8	30.4	60.4	60.8	71.7	52.6	33.4	30.7	49.73	1,237.1	—	632.22
Pythia2026.04		48.7	24.8	56.8	58.1	66.7	41.6	26.1	26.2	43.63	360.9	—	197.71
OPT2026.04		47.5	22.2	55.3	57.2	66.1	40.7	25.7	26.6	42.68	360.8	—	197.57
BiSpikCLM2026.04		46.3	24.3	55.6	56.8	63.4	41.7	25.2	24.3	42.19	131.9	0.184	66.89
TriSpikCLM2026.04		45.8	24.5	55.6	56.1	63.4	41.3	25.5	24.8	42.12	141.6	0.411	141.21
BiSpikCLM2026.04		45.7	23.5	54.2	56.3	62.3	40.2	24.5	24	41.33	66.7	0.192	37.16
TriSpikCLM2026.04		44.5	23.7	54.2	55.3	62.3	40.4	24.6	23.6	41.08	74.2	0.426	74.32
Pythia2026.04		43.7	19.8	52.8	55.1	62.7	33.6	20.1	24.2	39	125.7	—	126.01
OPT2026.04		43.6	19.3	52.3	54.6	62.4	32.1	20.2	23.7	38.6	125.6	—	125.95
BiSpikCLM2026.04		42.1	21.4	52.1	56.1	60.5	33.1	21.9	23.5	38.84	84.2	0.178	16.75
TriSpikCLM2026.04		41.8	21.9	52.8	55.7	60.4	34	21.3	23.1	38.87	88.3	0.377	35.35
BiSpikCLM2026.04		41.5	21.7	52.3	55.1	59.7	32.7	21.2	23.8	38.48	43.4	0.182	9.31
TriSpikCLM2026.04		41.5	21.7	52.4	54.9	59.1	31.6	20.8	23.1	38.14	47.7	0.404	18.61
BiSpikCLM2026.04		39.4	19	51.2	53	57.5	29.2	19.7	23.1	36.5	23.1	0.173	9.43
BiSpikCLM2026.04		39.1	18.9	50.3	52.7	56.7	28.1	19.8	22.9	36.05	12.1	0.196	5.24
TriSpikCLM2026.04		38.9	18.5	51.5	52.9	58	28.3	19.2	22.9	36.27	25.8	0.386	19.92
TriSpikCLM2026.04		38.5	18.3	51.3	52.3	57.7	29.1	19.2	22.5	36.11	13.7	0.412	10.74
SpikeGPT2026.04		35.2	17.7	50.7	47.3	55.1	27.6	17.3	23.1	34.25	18.3	0.168	16.53
SpikeGPT2026.04		32.3	16.2	50.2	45.7	54.6	25.3	15.7	20.6	32.58	3.66	0.174	3.29