LLaVA 1.6 Mistral 7B: benchmark results
Provider: LLaVA Team.
Aggregate Rating 1314 ± 48, rank #8458 of 10916 rated models, from 17 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| ShellfishNet Captioning | 85.83 | BERTScore F1 (0-100; the paper's 0-1 score times 100) of the | 33.3 |
| IDEAL-Bench - Rotation Accuracy | 19.2 | Rotation accuracy (%): share of matched objects in the 7 non | 20.8 |
| CharXiv-D | 35.4 | Score (%) | 17.8 |
| CharXiv-R | 13.9 | Validation chart-visible reasoning accuracy (%) | 8.5 |
| IDEAL-Bench - Grid Coverage | 8 | Grid coverage rate (%): share of ground-truth grid cells (cl | 8.3 |
| IDEAL-Bench - Position Accuracy | 2.4 | Position accuracy (%): share of matched objects in the 7 non | 8.3 |
| ExChart-Bench | 70.52 | Adaptive MAPE (%; 33,757 values in 3,600 unlabeled charts, p | 0 |
| ExChart-Bench - Bar Charts | 68.96 | Adaptive MAPE (%; 9,500 values in 800 charts, per-value erro | 0 |
| ExChart-Bench - Line Charts | 72.64 | Adaptive MAPE (%; 9,494 values in 700 charts, per-value erro | 0 |
| ExChart-Bench - Pie Charts | 61.73 | Adaptive MAPE (%; 3,055 values in 700 charts, per-value erro | 0 |
| ExChart-Bench - Radar Charts | 69.4 | Adaptive MAPE (%; 3,294 values in 700 charts, per-value erro | 0 |
| ExChart-Bench - Scatter Plots | 73.53 | Adaptive MAPE (%; 8,414 values in 700 charts, per-value erro | 0 |
Interactive version: theaggregate.ai/model?slug=llava-1-6-mistral-7b · How It Works · Data refreshed daily, snapshot 2026-11-08.