Cambrian-13B: benchmark results
Provider: NYU. Released 2024-06-24. Access: Open.
Aggregate Rating 1373 ± 33, rank #7083 of 10916 rated models, from 244 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| OpenVLM SEEDBench IMG - Instance Interaction | 78.4 | Accuracy (%) | 91.8 |
| OpenVLM MME - Celebrity | 161.5 | Score | 84 |
| OpenVLM A-Bench - Aesthetic Quality | 65.8 | Accuracy (%) | 83.3 |
| OpenVLM AI2D - Eclipses | 93.5 | Accuracy (%) | 80.4 |
| OpenVLM VCR - English Hard | 30.1 | Jaccard (%) | 75.5 |
| OpenVLM MMMU - Music | 36.7 | Accuracy (%) | 75.4 |
| OpenVLM Q-Bench - Yes-or-No (Distortion) | 84 | Accuracy (%) | 73.7 |
| OpenVLM MMMU - Materials | 40 | Accuracy (%) | 71.7 |
| OpenVLM MMBench V1.1 EN - Identity Reasoning | 97.4 | Accuracy (%) | 71 |
| OpenVLM MMBench V1.1 CN - Action Recognition | 87.2 | Accuracy (%) | 68.7 |
| OpenVLM MathVista - LOG | 24.3 | Accuracy (%) | 67.2 |
| OpenVLM LLaVA-Bench - Conv | 79.9 | Score | 64.8 |
Interactive version: theaggregate.ai/model?slug=cambrian-13b · How It Works · Data refreshed daily, snapshot 2026-11-08.