List your company

Joint Audio-Video Generation on Identity-aware T2AV (test)

0.606AESIdentity-as-Presence
0.529040.549020.5690.58898Mar 18, 2026

Evaluation results

MethodLinks
Identity-as-PresenceResolution=720p, Duration=5s, Backbone=Dual-tower (Ovi-based)2026.030.6066.9440.330.7290.1880.5970.8670.1320.6677.0638.1490.335
R2I+UniAVGen I2AVResolution=720p, Duration=5s, Input Type=I2AV2026.030.5986.0910.3150.8880.3270.3210.6220.1070.4034.6310.3680.23
Stand-InResolution=720p, Duration=5s, Input Type=I2AV (via Qwen-Image first frame)2026.030.576—————0.1590.120.561———
R2I+Ovi I2AVResolution=720p, Duration=5s, Input Type=I2AV2026.030.5545.4950.3040.8930.203—0.6670.1110.3836.8448.5550.229
HunyuanCustomResolution=720p, Duration=5s, Input Type=I2AV (via Qwen-Image first frame)2026.030.553—————0.7110.1250.563———
R2I+LTX-2 I2AVResolution=720p, Duration=5s, Input Type=I2AV2026.030.5385.3020.2681.1310.205—0.7330.1120.3735.7778.9080.238
PhantomResolution=720p, Duration=5s, Input Type=I2AV (via Qwen-Image first frame)2026.030.532—————0.8440.1240.631———