VITA-1.5-7B: benchmark results
Provider: Tencent. Access: Open.
Aggregate Rating 1542 ± 33, rank #1814 of 10916 rated models, from 36 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| VSTAT - Set | 43.1 | Score (%) on questions about set state structures: 1,500 que | 72.7 |
| VSTAT - Attribute | 38.6 | Score (%) on questions about attribute-type state elements: | 63.6 |
| VSTAT - Location | 36.3 | Score (%) on questions about location-type state elements: 1 | 63.6 |
| VideoOdyssey-AV - Action Recognition | 30 | Multiple-choice accuracy (%) on the 60 action recognition qu | 54.5 |
| VideoOdyssey-AV - Object Reasoning | 30.4 | Multiple-choice accuracy (%) on the 56 object reasoning ques | 54.5 |
| VideoOdyssey-AV - Speech Content Reasoning | 30.7 | Multiple-choice accuracy (%) on the 62 speech content reason | 54.5 |
| VSTAT | 31.5 | Average score (%) over all questions: 1,500 questions on 834 | 50 |
| VSTAT - Sequence | 33 | Score (%) on questions about sequence state structures: 1,50 | 45.5 |
| VideoOdyssey-AV - Causal Reasoning | 24.1 | Multiple-choice accuracy (%) on the 58 causal reasoning ques | 45.5 |
| VSTAT - Atomic | 29.4 | Score (%) on questions about atomic state structures: 1,500 | 40.9 |
| VideoOdyssey-AV - OCR | 27.1 | Multiple-choice accuracy (%) on the 48 OCR questions (recogn | 40.9 |
| VideoOdyssey-AV - Object Recognition | 28.6 | Multiple-choice accuracy (%) on the 63 object recognition qu | 40.9 |
Interactive version: theaggregate.ai/model?slug=vita-1-5-7b · How It Works · Data refreshed daily, snapshot 2026-11-08.