VITA: benchmark results
Provider: Tencent. Released 2024-08-12. Access: Open.
Aggregate Rating 1446 ± 25, rank #3947 of 10916 rated models, from 193 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| AV-Odyssey - Action Tracing | 39.5 | Score (%) | 94.1 |
| AV-Odyssey - Film and Music Matching | 33.5 | Score (%) | 94.1 |
| AV-Odyssey - Intricacy | 36.8 | Score (%) | 91.2 |
| AV-Odyssey - Eating Sound Recognition | 29.5 | Score (%) | 85.3 |
| AV-Odyssey - Music Sentiment Analysis | 26.9 | Score (%) | 85.3 |
| AV-Odyssey - Action Prediction | 34.2 | Score (%) | 82.4 |
| AV-Odyssey - Hallucination | 31 | Score (%) | 82.4 |
| AV-Odyssey - Meme Understanding | 45 | Score (%) | 82.4 |
| AV-Odyssey - Tone | 26.4 | Score (%) | 76.5 |
| OpenVLM MMMU - Sociology | 66.7 | Accuracy (%) | 76.3 |
| CG-Bench - Open-Ended | 28 | Open-Ended Accuracy (%) | 76.2 |
| CG-Bench | 33 | Long-Video MCQ Accuracy (%) | 71.4 |
Interactive version: theaggregate.ai/model?slug=vita · How It Works · Data refreshed daily, snapshot 2026-11-08.