| Paper | Code | Accuracy | ModelName | ReleaseDate |
|---|---|---|---|---|
| LEAP: Learned Block-wise Evidence Retrieval for Long Audio-Video Perception | 46.5 | LEAP (extra data) | 2026-09-30 | |
| OmniReasoning: Pushing the Limits of Audio-Visual Joint Reasoning | ✓ Link | 44.7 | OmniReasoning-30B-A3B (extra data) | 2026-09-30 |
| OmniSeek: Native Tool Integration for Multi-turn Audio-Visual Reasoning | 44.2 | OmniSeek (ours) | 2026-10-01 | |
| Realtime-Venus: A full-duplex interaction system with asynchronous delegation | 38.9 | Realtime-Venus-Omni | 2026-09-12 | |
| LatentOmni: Rethinking Omni-Modal Understanding via Unified Audio-Visual Latent Reasoning | ✓ Link | 35.1 | LatentOmni | 2026-05-21 |