ResearchBenchmarksUnconditional Co-generation on Image-Text-AudioFor AI assistantsMarkdownContent indexUnconditional Co-generation on Image-Text-AudioFollow26.76CLIP Score (Text-Image)MUNI20.946422.455723.96525.4743Jun 15, 2026Evaluation resultsMethodMethodLinksCLIP Score (Text-Image)CLAP Score (Text-Audio)AIS Score (Audio-Image)MUNICategory=Generalists,...Category=Generalists, Aggregation=PoE2026.0626.7623.7381.24MUNI†Category=Generalists,...Category=Generalists, Aggregation=Hellinger2026.0626.7424.8382.8MoPoECategory=Multimodal VAEsCategory=Multimodal VAEs2026.0625.8321.4774.37MMVAECategory=Multimodal VAEsCategory=Multimodal VAEs2026.0625.5220.0676.68OmniFlowCategory=GeneralistsCategory=Generalists2026.0621.1714.2350.95