Palmyra Vision 003: benchmark results
Provider: Writer. Access: API.
Aggregate Rating 1589 ± 22, rank #1133 of 10916 rated models, from 154 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| HELM VHELM - Mmmu - Literature | 90 | Prefix Quasi-Exact Match (%) | 98.9 |
| HELM Image2Struct - Image2Latex Algorithm | 67.48 | EMS | 94.7 |
| HELM VHELM - Blink Perception | 83.35 | PEM | 91.5 |
| HELM VHELM - Visual Perception | 71.63 | Mean win rate | 88.3 |
| HELM VHELM - Image2Webpage | 62.59 | EMD-Sim | 88.2 |
| HELM VHELM - Blink Perception - Relative Reflectance | 67.91 | Prefix Quasi-Exact Match (%) | 87.2 |
| HELM VHELM - Crossmodal 3600 - Swahili Locale | 3.48 | Prometheus Vision Rating (1-5) | 87.2 |
| HELM VHELM - Mm Star Reasoning - Math | 81.6 | Prefix Quasi-Exact Match (%) | 85.1 |
| HELM Image2Struct - Image2Struct Wild | 70.39 | Mean win rate | 84.2 |
| HELM VHELM - Mmmu - History | 83.33 | Prefix Quasi-Exact Match (%) | 84 |
| HELM VHELM - Vqa Base | 85.3 | PEM | 83 |
| HELM VHELM - Crossmodal 3600 - Hindi Locale | 3.5 | Prometheus Vision Rating (1-5) | 81.9 |
Interactive version: theaggregate.ai/model?slug=palmyra-vision-003 · How It Works · Data refreshed daily, snapshot 2026-11-08.