Detailed Caption on KVBench Detailed Caption en benchmark leaderboard · SOTA2 Research

Detailed Caption on KVBench Detailed Caption en

62.3Overall ScoreGPT-Image

Evaluation Results

Method	Links
GPT-Image2026.04		62.3	68.77	66.36	54.76	48.5	66.21	73.18
gemini-2.5-flash2026.04		61.23	75.15	74.54	38.95	62.24	67.96	73.54
Seedream-4.02026.04		59.67	67.03	69.1	49.92	45.08	66.05	70.82
gemini-3-pro-image2026.04		57.55	77	79.47	15.11	60.41	53.11	76.2
FLUX.2-dev2026.04		56.64	53.15	62.78	46.84	52.94	58.56	65.54
FLUX.2-max2026.04		50.18	54.68	55.74	29.37	43.62	52.91	64.74
SD3.5-Large2026.04		39.35	31.15	45.6	34.32	37.04	39.97	48.96
Qwen-Image2026.04		37.24	33.47	50.47	33.38	34.4	23.13	48.67
FLUX.1-Krea-dev2026.04		36.65	29.77	43.62	29.32	25.38	44.48	47.74
HiDream2026.04		35.84	33	45.62	28.68	28.04	35.04	45.28
Z-Image2026.04		28.9	25.67	45.4	21.61	23.78	15.32	41.4
FLUX.1-dev2026.04		24.32	15.25	31.98	12.35	18.38	30.17	37.78
BAGEL-7B-MoT2026.04		20.31	11.8	32.36	12.04	10.7	25.59	29.38
Show-o22026.04		15.14	18.59	22.34	15.21	14.38	5.63	14.7
OmniGen22026.04		14.27	8.27	23.84	11.84	11.06	6.6	23.02
Janus-Pro2026.04		9.96	5.59	19.76	9.6	5.04	4.36	10.38