InternVL2.5-1B-MPO: benchmark results
Provider: Shanghai AI Lab. Released 2024-12-20. Access: Open.
Aggregate Rating 1355 ± 22, rank #7581 of 10916 rated models, from 489 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| OpenVLM MMT-Bench - Count | 61.9 | Score (%) | 99.3 |
| OpenVLM MMT-Bench - Counting By Visual Prompting | 60 | Score (%) | 99.3 |
| OpenVLM CCBench - Food and Clothes | 84.3 | Accuracy (%) | 98.2 |
| OpenVLM MMT-Bench - Temporal Anticipation | 85 | Score (%) | 97.8 |
| OpenVLM CCBench - Traditional Shows | 83.3 | Accuracy (%) | 96.5 |
| CCBench | 78.2 | Accuracy (%) | 95.2 |
| OpenVLM MMT-Bench - Micro Expression Recognition | 45 | Score (%) | 94.9 |
| OpenVLM POPE | 89.4 | Overall (%) | 93 |
| OpenVLM MMT-Bench - Crowd Counting | 70 | Score (%) | 92.7 |
| OpenVLM MMT-Bench - Counting By Reasoning | 90 | Score (%) | 92.2 |
| OpenVLM MMT-Bench - Temporal Ordering | 40 | Score (%) | 92 |
| OpenVLM CCBench - Calligraphy and Painting | 84.2 | Accuracy (%) | 91.9 |
Interactive version: theaggregate.ai/model?slug=internvl2-5-1b-mpo · How It Works · Data refreshed daily, snapshot 2026-11-08.