MAmmoTH-VL-8B: benchmark results
Provider: CMU.
Aggregate Rating 1406 ± 37, rank #5779 of 10916 rated models, from 13 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| MMMU Benchmark | 50.8 | Validation Score | 44.9 |
| KOFFVQA - Recognition | 60 | Score (%) | 40.7 |
| KOFFVQA - Graph and Chart Understanding | 42 | Score (%) | 23.5 |
| KOFFVQA - Document Understanding | 24.33 | Score (%) | 17.9 |
| Video-MMMU | 41.78 | Overall Score | 15.4 |
| KOFFVQA - Korean OCR | 5 | Score (%) | 9.3 |
| KOFFVQA - Korean Recognition | 0 | Score (%) | 8.6 |
| KOFFVQA - Object Attributes | 36.33 | Score (%) | 6.2 |
| KOFFVQA - Commonsense Reasoning | 33.11 | Score (%) | 4.9 |
| KOFFVQA | 25.96 | Overall Score (%) | 3.7 |
| KOFFVQA - Relationship | 17.33 | Score (%) | 2.5 |
| KOFFVQA - Table Understanding | 5.33 | Score (%) | 2.5 |
Interactive version: theaggregate.ai/model?slug=mammoth-vl-8b · How It Works · Data refreshed daily, snapshot 2026-11-08.