Ovis2-34B: benchmark results
Provider: Alibaba. Released 2025-02-21. Access: Open.
Aggregate Rating 1594 ± 12, rank #1074 of 10916 rated models, from 589 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| Open LMM Reasoning - WeMath - Understanding of Solid Figures | 75.5 | Accuracy (%) | 100 |
| OpenVLM MMBench V1.1 CN - Physical Property Reasoning | 86.1 | Accuracy (%) | 100 |
| OpenVLM MMT-Bench - IEJ | 77.5 | Score (%) | 100 |
| OpenVLM MMT-Bench - Image Quality Assessment | 90 | Score (%) | 100 |
| OpenVLM MTVQA - Russian | 16.8 | Accuracy (%) | 100 |
| OpenVLM MMBench V1.1 CN - AR | 93.1 | Accuracy (%) | 99.8 |
| OpenVLM MME - Numerical Calculation | 192.5 | Score | 99.8 |
| OpenVLM MMT-Bench - Equation-to-LaTeX | 100 | Score (%) | 99.8 |
| OpenVLM MMT-Bench - MemU | 92.5 | Score (%) | 99.8 |
| OpenVLM MMT-Bench - Visual Prompt Understanding | 95 | Score (%) | 99.8 |
| OpenVLM MMBench V1.0 CN - AR | 89.2 | Accuracy (%) | 99.6 |
| OpenVLM MME - Cognition | 742.9 | Score | 99.6 |
Interactive version: theaggregate.ai/model?slug=ovis2-34b · How It Works · Data refreshed daily, snapshot 2026-11-08.