VideoChat2-HD: benchmark results
Provider: Other.
Aggregate Rating 1496 ± 42, rank #2420 of 10916 rated models, from 81 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| MLVU-Test - Generation | 3.99 | Generation score, GPT-4 judged (0-10) | 84.4 |
| OpenVLM Video - MVBench - Action Antonym | 89 | Accuracy (%) | 77.8 |
| OpenVLM Video - MVBench - Egocentric Navigation | 38.5 | Accuracy (%) | 76.4 |
| OpenVLM Video - MVBench - Action Sequence | 77.5 | Accuracy (%) | 75 |
| OpenVLM Video - MVBench - Object Existence | 93 | Accuracy (%) | 75 |
| OpenVLM Video - MVBench - Counterfactual Inference | 73 | Accuracy (%) | 72.2 |
| MLVU-Test - Action Count | 23.3 | Multiple-choice accuracy (%) | 69.1 |
| MLVU-Dev - Generation | 3.99 | Generation score, GPT-4 judged (0-10) | 65.5 |
| OpenVLM Video - MVBench - Moving Count | 69.5 | Accuracy (%) | 65.3 |
| OpenVLM Video - MVBench - Moving Direction | 45 | Accuracy (%) | 63.9 |
| OpenVLM Video - MVBench - Moving Attribute | 90.5 | Accuracy (%) | 61.1 |
| OpenVLM Video - MVBench - Fine-grained Action | 48 | Accuracy (%) | 59.7 |
Interactive version: theaggregate.ai/model?slug=videochat2-hd · How It Works · Data refreshed daily, snapshot 2026-11-08.