Text-to-Image Generation on UniGenBench (Granular Evaluation) benchmark leaderboard · SOTA2 Research

Text-to-Image Generation on UniGenBench (Granular Evaluation)

91.77Attribute ScoreLLaDA2.0-Uni

Evaluation Results

Method	Links
LLaDA2.0-Uni2026.04		91.77	95.3	93.67	85.65	86.42	63.99	72.19	85.82	90.3	31.23	79.63	—	—
Qwen-Image2026.04		87.93	94.7	94.15	82.6	80.08	51.59	60.96	72.94	86.57	72.13	78.36	—	—
CapRL-T2I2026.06		85.04	—	—	79.18	—	61.24	—	—	—	—	80.86	—	57.39
Seedream 3.02026.04		84.62	98.19	94.9	83.14	80.18	51.83	60.3	72.32	88.74	69.86	78.41	—	—
FLUX2 Klein 9B2026.06		82.91	—	—	76.52	—	55.05	—	—	—	—	78.59	—	55.33
Z-Image2026.04		82.48	96.8	94.46	78.9	80.2	49.08	68.98	76.8	84.89	68.39	78.1	—	—
Lumina-DiMOO2026.04		81.62	89.7	90.03	71.12	78.43	45.45	70.45	73.32	82.84	25.57	71.12	—	—
CapRL-T2I2026.06		79.49	—	—	75.38	—	44.5	—	—	—	—	71.77	—	50.29
Pave-GRPO (F_mar_sde)2026.06		72.22	85.4	90.66	67.11	71.95	—	60.56	52.84	73.51	35.06	63.96	30.28	—
OmniGen22026.04		72.12	91.9	86.39	62.83	68.27	32.5	59.89	56.31	71.64	29.02	63.09	—	—
Pave-GRPO (F_snr_sde)2026.06		70.51	86.4	89.87	65.68	71.45	—	60.56	50.26	73.88	30.46	63.23	33.26	—
Flow-GRPO2026.06		70.19	85.8	89.08	66.06	69.67	—	60.7	51.03	73.88	30.75	62.65	29.36	—
Flow-CPS2026.06		69.66	86.1	90.03	65.87	71.32	—	61.36	52.19	73.51	27.3	62.65	29.13	—
DanceGRPO2026.06		68.48	85	88.29	61.12	70.05	—	59.36	48.2	71.64	30.46	61.15	28.9	—
FLUX.1 [dev]2026.04		67.84	83.9	88.92	62.17	67.26	30.91	60.96	47.04	71.83	32.18	61.3	—	—
Janus-Pro2026.04		67.74	90.8	86.71	64.26	68.4	37.05	64.44	62.11	72.01	2.59	61.61	—	—
BAGEL2026.04		67.74	90.2	85.6	61.98	70.69	30.23	66.44	58.12	76.49	7.76	61.53	—	—
Flux.1-dev2026.06		64.32	84.6	85.76	61.22	66.37	—	58.69	46.26	70.71	33.62	59.86	27.06	—
FLUX.1-dev2026.06		64.32	—	—	61.12	—	27.52	—	—	—	—	59.92	—	34.48
Emu3-Gen2026.04		51.39	86.8	77.06	40.11	49.75	19.32	52.94	36.86	44.78	1.15	46.02	—	—
MMaDA2026.04		48.39	82.4	56.65	37.83	50.25	17.95	55.75	32.35	30.22	1.15	41.35	—	—