PandaGPT-13B: benchmark results
Provider: UCSD. Released 2023-05-25.
Aggregate Rating 1209 ± 52, rank #9366 of 10489 rated models, from 422 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| OpenVLM MMT-Bench - Gaze Estimation | 40 | Score (%) | 99 |
| OpenVLM MMT-Bench - Multiple View Image Understanding | 50 | Score (%) | 98.1 |
| OpenVLM MMT-Bench - Font Recognition | 35 | Score (%) | 88.1 |
| OpenVLM MMT-Bench - Order Hallucination | 50 | Score (%) | 81.8 |
| OpenVLM MMT-Bench - Micro Expression Recognition | 35 | Score (%) | 79.4 |
| OpenVLM MMT-Bench - Web Shopping | 45 | Score (%) | 78.9 |
| OpenVLM MMT-Bench - Artwork Emotion Recognition | 55 | Score (%) | 74.8 |
| OpenVLM MMT-Bench - Image Quality Assessment | 50 | Score (%) | 72.6 |
| OpenVLM MMT-Bench - GUI Install | 45 | Score (%) | 71.6 |
| OpenVLM MMT-Bench - Action Quality Assessment | 25 | Score (%) | 69.7 |
| OpenVLM MMT-Bench - Handwritten Retrieval | 30 | Score (%) | 67.2 |
| AVI-Bench - Visual-Referenced Audio Retrieval | 23.7 | Task score (%) on visual-referenced audio retrieval (264 que | 64 |
Interactive version: theaggregate.ai/model?slug=pandagpt-13b · How It Works · Data refreshed daily, snapshot 2026-11-20.