InternVideo2.5-Chat-8B: benchmark results
Provider: Other.
Aggregate Rating 1519 ± 25, rank #2081 of 10916 rated models, from 49 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| KPM-QA - Camera | 77.42 | Accuracy (%) by option matching on KPM-QA camera (camera mov | 100 |
| KPM-QA - Emotion | 92.21 | Accuracy (%) by option matching on KPM-QA emotion (emotions | 100 |
| KPM-QA - Movement | 83.7 | Accuracy (%) by option matching on KPM-QA movement (movement | 100 |
| FAQ (Deepfake Forensics) - Facial Perception | 54.6 | Accuracy (%) on FAQ's Level 1 facial-perception questions: f | 83.3 |
| HUG-VIS - Emotion Recognition (Video + Text) | 78.77 | Seven-class emotion recognition accuracy (%; 8,400 seated ha | 83.3 |
| HUG-VIS - Emotion Recognition (Video) | 46.42 | Seven-class emotion recognition accuracy (%; 8,400 seated ha | 83.3 |
| HAVEN - Keyframe Selection | 27.3 | Keyframe selection F1 (%) against human-annotated keyframes | 66.7 |
| KPM-QA - Environment | 94.22 | Accuracy (%) by option matching on KPM-QA environment (indoo | 66.7 |
| HAVEN - Saliency Ranking | 0.17 | Shot saliency ranking: Spearman rank correlation (-1 to 1) b | 60 |
| HAVEN - Saliency Ranking (With Document) | 0.18 | Shot saliency ranking: Spearman rank correlation (-1 to 1) b | 60 |
| KPM-QA - Human | 75.74 | Accuracy (%) by option matching on KPM-QA human (how many pe | 50 |
| LifeEval (Multiple-Choice) - Dynamic Task Reasoning | 58.38 | Accuracy (%) on the 370 multiple-choice Dynamic Task Reasoni | 48 |
Interactive version: theaggregate.ai/model?slug=internvideo2-5-chat-8b · How It Works · Data refreshed daily, snapshot 2026-11-08.