ResearchBenchmarksMany-to-one Image Generation ((T+A) → I) on Triplet (text, image, audio)Many-to-one Image Generation ((T+A) → I) on Triplet (text, image, audio)Follow25.57CLIP Score (T→I)FlowBind23.999624.407324.81525.2227Dec 17, 2025Evaluation resultsMethodMethodLinksCLIP Score (T→I)AIS Score (A→I)FlowBind2025.1225.5757.93CoDi2025.1225.1757.52OmniFlow2025.1224.0654.9