SAIL-VL-1.5-2B: benchmark results
Provider: ByteDance. Released 2024-12-01. Access: Open.
Aggregate Rating 1459 ± 18, rank #3387 of 10916 rated models, from 238 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| Open LMM Reasoning - LogicVista - spatial | 34.6 | Accuracy (%) | 98.1 |
| OpenVLM MMBench V1.1 EN - Attribute Recognition | 97.8 | Accuracy (%) | 98.1 |
| OpenVLM OCRBench - Text Recognition | 93.33 | Correct Answers (%) | 94.7 |
| OpenVLM OCRBench - Handwritten Mathematical Expression Recognition | 81 | Correct Answers (%) | 94.5 |
| OpenVLM AI2D - Photosynthesis and Respiration | 88.6 | Accuracy (%) | 94.2 |
| OpenVLM OCRBench - Scene Text-centric VQA | 94 | Correct Answers (%) | 93.7 |
| OpenVLM AI2D - Eclipses | 96.8 | Accuracy (%) | 93.1 |
| OpenVLM MMBench V1.1 CN - Attribute Recognition | 95.5 | Accuracy (%) | 92.8 |
| OpenVLM OCRBench | 89.1 | Score (normalized) | 91.2 |
| OpenVLM MMStar - Coarse Perception | 72.4 | Accuracy (%) | 89.9 |
| OpenVLM MathVista - MWP | 80.6 | Accuracy (%) | 86.9 |
| OpenVLM AI2D - Rock Cycle | 73.1 | Accuracy (%) | 85.9 |
Interactive version: theaggregate.ai/model?slug=sail-vl-1-5-2b · How It Works · Data refreshed daily, snapshot 2026-11-08.