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.
Flux1.1 [pro]: Advanced text-to-image generation
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CV 1years
2025 1verdicts
CONDITIONAL 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.