Video-R1: benchmark results
Provider: Other.
Aggregate Rating 1572 ± 28, rank #1369 of 10916 rated models, from 58 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| TOC-Bench - Reappearance Identity | 100 | Reappearance Identity accuracy (%) on TOC-Bench (2,323 human | 90 |
| MuseBench - Stage Performing Arts (Single-Select) | 38.49 | Stage-performing-arts chance-adjusted accuracy (%): raw sing | 70.4 |
| MuseBench - Game Arts (Single-Select) | 13.87 | Game-arts chance-adjusted accuracy (%): raw single-select ac | 66.7 |
| VideoZeroBench - Spatial Grounding | 5.5 | Mean visual IoU (%) of the bounding boxes the model predicts | 62.5 |
| VideoOdyssey-V - Emotional Reasoning | 39.4 | Multiple-choice accuracy (%) on the 71 emotional reasoning q | 61.8 |
| MMVIAD - Anomaly Detection | 41.4 | Accuracy (%; MMVIAD-Standard test split, 1,101 rendered 2-se | 60 |
| MMVIAD - Unseen Categories | 46.5 | Mean of three accuracies and visible-time mIoU (%; MMVIAD-Un | 60 |
| MuseBench - Single-Select (Overall) | 28.41 | Overall chance-adjusted accuracy (%): raw single-select accu | 59.3 |
| MuseBench - Stage Performing Arts (Multi-Select) | 10.26 | Stage-performing-arts exact-match accuracy (%) on multi-sele | 57.4 |
| MuseBench: Benchmarking Intent-Level Audiovisual Arts Understanding in MLLMs | 26.73 | Overall ACC (%) | 55.6 |
| Video-Holmes | 36.5 | Overall (imported) | 53.3 |
| VideoOdyssey-V - Spatial Reasoning | 36.1 | Multiple-choice accuracy (%) on the 61 spatial reasoning que | 52.9 |
Interactive version: theaggregate.ai/model?slug=video-r1 · How It Works · Data refreshed daily, snapshot 2026-11-08.