LLaVA-CoT: benchmark results
Provider: LLaVA Team. Released 2024-11-21. Access: Open.
Aggregate Rating 1463 ± 19, rank #3122 of 10916 rated models, from 270 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| OpenVLM MMMU - Design | 86.7 | Accuracy (%) | 95.8 |
| OpenVLM MMMU - Art | 80 | Accuracy (%) | 92.8 |
| OpenVLM MMBench V1.1 EN - Identity Reasoning | 98.7 | Accuracy (%) | 91.5 |
| Open LMM Reasoning - WeMath - InadequateGeneralization (Loose) | 14.7 | Accuracy (%) | 87.7 |
| OpenVLM MMMU - Electronics | 76.7 | Accuracy (%) | 85.5 |
| OpenVLM MMMU - Geography | 60 | Accuracy (%) | 83.7 |
| OpenVLM MMBench V1.1 EN - Celebrity Recognition | 89.1 | Accuracy (%) | 83.6 |
| Open LMM Reasoning - WeMath - RoteMemorization (Strict) | 69.6 | Accuracy (%) | 82.7 |
| OpenVLM MMMU - Chemistry | 56.7 | Accuracy (%) | 82 |
| OpenVLM MMMU - Architecture and Engineering | 43.3 | Accuracy (%) | 81.6 |
| Open LMM Reasoning - WeMath - RoteMemorization (Loose) | 12.6 | Accuracy (%) | 81.5 |
| OpenVLM MMMU - Finance | 63.3 | Accuracy (%) | 80.9 |
Interactive version: theaggregate.ai/model?slug=llava-cot · How It Works · Data refreshed daily, snapshot 2026-11-08.