Image Classification on ImageNet and OOD variants 1.0 (test) benchmark leaderboard · SOTA2 Research

Image Classification on ImageNet and OOD variants 1.0 (test)

62.4ImageNet-A AccuracyCoOp + Ours

Evaluation Results

Method	Links
CoOp + Ours2026.04		62.4	74.03	67.5	81.02	52.58	67.51	65.88
MaPLe + DynaPrompt2026.04		60.72	72.71	66.34	79.57	50.25	65.92	64.22
CoOp + DynaPrompt2026.04		60.55	74.08	67.25	79.15	50.28	66.26	64.31
Ours2026.04		59.65	72.21	65.35	80.25	51.24	65.74	64.12
MaPLe + Ours2026.04		59.65	72.31	66.84	80.5	52.11	66.28	64.78
MaPLe + PromptAlign2026.04		59.37	—	65.29	79.33	50.23	—	63.56
MaPLe + TPT2026.04		58.08	71.87	64.87	78.12	48.16	64.22	62.31
CoOp + TPT2026.04		57.95	73.61	66.83	77.27	49.29	64.99	62.84
DynaPrompt2026.04		56.17	69.61	64.67	78.17	48.22	63.37	61.81
DiffTPT2026.04		55.68	70.3	65.1	75	46.8	62.26	60.65
TPT2026.04		54.77	68.98	63.45	77.06	47.06	62.06	60.81
C-TPT2026.04		52.9	69.3	63.4	78	48.5	62.42	60.7
CoCoOp2026.04		50.63	71.02	64.47	76.18	48.75	62.13	59.91
Any-shift Prompt2026.04		49.8	—	64.53	77.56	51.52	—	60.85
CoOp2026.04		49.71	71.51	64.2	75.21	47.99	61.72	59.28
CoPrompt2026.04		49.43	70.8	64.25	77.51	50.5	62.5	60.42
MaPLe2026.04		49.15	70.72	64.07	76.98	50.9	62.36	60.28
CLIP-ViT-B/162026.04		47.87	66.73	60.86	73.98	46.09	59.11	57.2