VideoLLaMA2.1-7B-16F: benchmark results
Provider: Other.
Aggregate Rating 1524 ± 46, rank #1924 of 10552 rated models, from 18 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| CoreCognition Single Image Intuitivephysics | 59.09 | Accuracy (%) | 87.5 |
| CoreCognition Single Image Continuity | 66.28 | Accuracy (%) | 84.4 |
| CoreCognition Single Image Toolusing | 98.96 | Accuracy (%) | 81.2 |
| CoreCognition Single Image Perspective | 38.07 | Accuracy (%) | 71.9 |
| CoreCognition Single Image Perceptualconstancy | 65.12 | Accuracy (%) | 65.6 |
| CoreCognition Single Image Hierarchy | 55.99 | Accuracy (%) | 62.5 |
| CoreCognition Single Image Intentionality | 72.2 | Accuracy (%) | 59.4 |
| CoreCognition Single Image Mean | 57.46 | Accuracy (%) | 59.4 |
| AffectSim (Reference View) | 16.25 | Macro-F1 (%; P-Ref: an informative camera trajectory chosen | 47.8 |
| CoreCognition Single Image Conservation | 47.45 | Accuracy (%) | 46.9 |
| CoreCognition Single Image Mechanicalreasoning | 52.06 | Accuracy (%) | 46.9 |
| AffectSim (Active Observation) | 11.2 | Macro-F1 (%; A-Obs: the view reached by the authors' two-sta | 43.5 |
Interactive version: theaggregate.ai/model?slug=videollama2-1-7b-16f · How It Works · Data refreshed daily, snapshot 2026-11-16.