MMAlaya2: benchmark results
Provider: DataCanvas. Released 2024-08-27. Access: Open.
Aggregate Rating 1454 ± 36, rank #3614 of 10916 rated models, from 438 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| OpenVLM MMT-Bench - Helmet Anomaly Detection | 95 | Score (%) | 97.1 |
| OpenVLM MME - Position | 170 | Score | 96.8 |
| OpenVLM MMBench V1.1 EN - Action Recognition | 93.2 | Accuracy (%) | 94.5 |
| OpenVLM MMT-Bench - Weather Recognition | 100 | Score (%) | 94.4 |
| OpenVLM AI2D - Volcano | 100 | Accuracy (%) | 92.8 |
| OpenVLM MMBench V1.1 CN - Action Recognition | 92.3 | Accuracy (%) | 92.6 |
| OpenVLM MMT-Bench - Food Recognition | 100 | Score (%) | 92.5 |
| OpenVLM SEEDBench IMG - Instance Identity | 80.6 | Accuracy (%) | 92.4 |
| OpenVLM SEEDBench IMG - Instance Location | 72.9 | Accuracy (%) | 92.2 |
| OpenVLM ChartQA Test - Augmented | 94.2 | Relaxed Accuracy (%) | 92.1 |
| OpenVLM MMT-Bench - Artwork Emotion Recognition | 60 | Score (%) | 91.7 |
| OpenVLM TextVQA Val | 80.5 | VQA Score (%) | 91.6 |
Interactive version: theaggregate.ai/model?slug=mmalaya2 · How It Works · Data refreshed daily, snapshot 2026-11-08.