Text-to-Image Generation on MS-COCO subset of 50 prompts 2015 benchmark leaderboard · SOTA2 Research

Text-to-Image Generation on MS-COCO subset of 50 prompts 2015

3.34VendiSemantic Browsing

Evaluation Results

Method	Links
Semantic Browsing2026.06		3.34	0.61	6.52	0.9
CADS2026.06		3.29	0.67	6.3	0.89
Guidance Interval.2026.06		2.96	0.71	6.42	0.92
High-Temp. Post-Hoc Opt.2026.06		2.85	0.66	6.56	0.92
Post-Hoc Diversity Opt.2026.06		2.79	0.67	6.51	0.92
Power-Law CFG2026.06		2.75	0.74	6.28	0.93
Stochastic VLM Seeding2026.06		2.6	0.76	6.53	0.93