AI RESEARCH
Uni-MMMU: A Massive Multi-discipline Multimodal Unified Benchmark
arXiv CS.CV
•
ArXi:2510.13759v3 Announce Type: replace Unified multimodal models aim to jointly enable visual understanding and generation, yet current benchmarks rarely examine their true integration. Existing evaluations either treat the two abilities in isolation or overlook tasks that inherently couple them. To address this gap, we present Uni-MMMU, a comprehensive and discipline-aware benchmark that systematically unfolds the bidirectional synergy between generation and understanding across eight reasoning-centric domains, including science, coding, mathematics, and puzzles.