Phantom: Subject-consistent video generation via cross-modal alignment

Benchmark Model Rank Results
open-domain-subject-to-video-on-opens2v-evalPhantom-Wan-14B#3Total Score: 0.5232Venue: Open-SourceAesthetics: 0.4639
open-domain-subject-to-video-on-opens2v-evalPhantom-Wan-1.3B#4Total Score: 0.5071Venue: Open-SourceAesthetics: 0.4667