MiniGPT-v2: benchmark results
Provider: Other.
Aggregate Rating 1420 ± 52, rank #5148 of 10916 rated models, from 44 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| ChartBench - Value Extraction | 22 | Acc+ (%) | 70 |
| VISTA (Video Grounding) - Object-Object | 43.1 | Mean spatio-temporal IoU (m_vIoU, %): per-frame box IoU summ | 66.7 |
| ChartBench - Global Conception | 18.1 | Acc+ (%) | 60 |
| EIBench | 52.89 | Recall | 58.3 |
| EIBench (complex) | 35.1 | Recall | 58.3 |
| UAVBench - Visual Grounding (RefDrone) | 12.46 | Visual grounding accuracy at IoU 0.5 (%) for a referring exp | 55.6 |
| VISTA (Video Grounding) - Animal-Object | 47.4 | Mean spatio-temporal IoU (m_vIoU, %): per-frame box IoU summ | 55.6 |
| VISTA (Video Grounding) - Freeform Queries | 45.13 | Mean spatio-temporal IoU (m_vIoU, %): per-frame box IoU summ | 55.6 |
| VISTA (Video Grounding) - Human-Human | 46 | Mean spatio-temporal IoU (m_vIoU, %): per-frame box IoU summ | 55.6 |
| VISTA (Video Grounding) - No Interaction | 44.3 | Mean spatio-temporal IoU (m_vIoU, %): per-frame box IoU summ | 55.6 |
| VISTA (Video Grounding) - Referral Queries | 46.62 | Mean spatio-temporal IoU (m_vIoU, %): per-frame box IoU summ | 55.6 |
| VISTA (Video Grounding) - Spatial | 43.1 | Mean spatio-temporal IoU (m_vIoU, %): per-frame box IoU summ | 55.6 |
Interactive version: theaggregate.ai/model?slug=minigpt-v2 · How It Works · Data refreshed daily, snapshot 2026-11-08.