A 57-task benchmark for multimodal image generation that uses automated visual parsers and an LLM judge to show GPT-4o-Native leads current models.
Evolving storytelling: benchmarks and methods for new character customization with diffusion models
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
citation-role summary
background 1
citation-polarity summary
fields
cs.CV 1years
2025 1verdicts
CONDITIONAL 1roles
background 1polarities
unclear 1representative citing papers
citing papers explorer
-
OmniGenBench: A Benchmark for Omnipotent Multimodal Generation across 50+ Tasks
A 57-task benchmark for multimodal image generation that uses automated visual parsers and an LLM judge to show GPT-4o-Native leads current models.