Image Classification on ImageNet (Standard Model Performance) benchmark leaderboard · SOTA2 Research

Image Classification on ImageNet (Standard Model Performance)

81.39Top-1 Accuracy (ViT-S)Full-Precision

Evaluation Results

Method	Links
Full-Precision2025.11		81.39	84.54	72.21	79.85	81.8	81.35	83.23	85.27
I-ViT2025.11		81.27	84.76	72.24	80.12	81.74	81.5	83.01	—
RepQ-ViT2025.11		81.19	84.39	72.03	79.68	81.77	81.19	83.05	85.12
PTQ4ViT2025.11		81	84.25	71.72	79.47	81.48	81.23	83.1	85.14
QUARK2025.11		80.72	83.78	71.29	79.4	81.52	81.06	82.81	85.03
I-BERT2025.11		80.47	83.7	71.33	79.11	80.79	80.15	81.86	—
RepQ-ViT2025.11		80.43	83.62	70.76	78.9	81.27	80.89	82.79	84.57
QUARK2025.11		80.12	83.28	70.29	78.6	81.42	81.08	82.93	84.78
FQ-ViT2025.11		78.68	82.76	70.92	78.44	81.12	79.97	82.38	82.53
PTQ4ViT2025.11		78.63	81.65	69.68	76.28	80.25	80.47	82.38	84.01
QUARK2025.11		68.84	74.56	60.68	72.83	78.76	77.85	81.44	83.12
RepQ-ViT2025.11		65.05	68.48	57.43	69.03	75.61	72.39	79.45	78.32
PTQ4ViT2025.11		42.57	30.69	36.96	34.08	64.39	73.48	76.09	74.02
FQ-ViT2025.11		4.26	0.1	58.66	45.51	64.63	48.63	66.5	52.09
FQ-ViT2025.11		0.1	0.1	0.1	0.1	0.1	0.1	0.1	0.1