VideoChat-R1.5-7B: benchmark results
Provider: Other. Access: Open.
Aggregate Rating 1580 ± 33, rank #1248 of 10916 rated models, from 45 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| TOC-Bench - Reappearance Identity | 100 | Reappearance Identity accuracy (%) on TOC-Bench (2,323 human | 90 |
| RepProbe Real - OBO | 19.4 | OBO (%) | 66.7 |
| MMVIAD - Object Classification | 66.2 | Accuracy (%; MMVIAD-Standard test split, 1,101 rendered 2-se | 60 |
| TOC-Bench - Conditional State | 40.6 | Conditional State accuracy (%) on TOC-Bench (2,323 human-ver | 60 |
| VideoZeroBench | 8.6 | Level-3 accuracy (%), standard end-to-end question answering | 56.2 |
| EG-VQA - Descriptive | 10.65 | Strict accuracy (%) on descriptive questions (directly obser | 50 |
| EG-VQA - Evidence Grounding | 11.86 | Evidence-grounded F1 (%, EG-F1 at temporal IoU 0.3 and seman | 50 |
| EG-VQA - Temporal | 8.12 | Strict accuracy (%) on temporal questions (event order and p | 50 |
| RepProbe Real - Normalized MAE | 1.4 | Normalized MAE (lower is better) | 50 |
| RepProbe Real - OBO Few | 25.7 | OBO Few (%) | 50 |
| RepProbe Synthetic - Exact | 5.2 | Exact (%) | 50 |
| RepProbe Synthetic - Normalized MAE | 0.72 | Normalized MAE (lower is better) | 50 |
Interactive version: theaggregate.ai/model?slug=videochat-r1-5-7b · How It Works · Data refreshed daily, snapshot 2026-11-08.