Doubao-1.5 Vision Pro: benchmark results
Provider: ByteDance.
Aggregate Rating 1598 ± 30, rank #1036 of 10916 rated models, from 30 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| VeriOCRBench - Visual Dimension (Assisted) | 79.25 | Diagnosis accuracy (%; the 400 trap-injected tasks of the vi | 100 |
| VeriOCRBench - Contextual Dimension (Assisted) | 97.5 | Diagnosis accuracy (%; the 400 trap-injected tasks of the co | 92.9 |
| VeriOCRBench - Logical Dimension (Assisted) | 75.25 | Diagnosis accuracy (%; the 400 trap-injected tasks of the lo | 92.9 |
| DrVD-Bench | 64.7 | Overall Accuracy (%) | 83.3 |
| RBench-V | 15.6 | Overall Score (%) | 74.1 |
| VeriOCRBench - Diagnosis Accuracy (Assisted) | 75.69 | Diagnosis accuracy (%; share of VeriOCRBench's 1,600 trap-in | 71.4 |
| AI for Education Visual Reasoning - match (process) | 44.4 | Accuracy (%) | 64.6 |
| AEGIS (Academic Image Forensics) - Textual Artifact Recognition | 72.32 | Macro-F1 (%) over two classes, telling whether the text regi | 60 |
| AEGIS (Academic Image Forensics) | 34.07 | Normalized Forensic Index (%): harmonic mean of the macro-F1 | 52 |
| AI for Education Visual Reasoning - odd one out | 48.1 | Accuracy (%) | 49.2 |
| AI for Education Visual Reasoning - pattern completion (linear) | 45.3 | Accuracy (%) | 48.5 |
| AI for Education Visual Reasoning - match (figure) | 42.1 | Accuracy (%) | 46.2 |
Interactive version: theaggregate.ai/model?slug=doubao-1-5-vision-pro · How It Works · Data refreshed daily, snapshot 2026-11-08.