SAIL-VL-1.5-8B: benchmark results
Provider: ByteDance. Released 2024-12-01. Access: Open.
Aggregate Rating 1513 ± 20, rank #2138 of 10916 rated models, from 233 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| OpenVLM AI2D - Eclipses | 100 | Accuracy (%) | 98.6 |
| OpenVLM OCRBench - Scene Text-centric VQA | 96 | Correct Answers (%) | 98.6 |
| OpenVLM AI2D - Parts of the Earth | 88.5 | Accuracy (%) | 98.4 |
| OpenVLM MMBench V1.1 EN - Image Emotion | 84.4 | Accuracy (%) | 98.2 |
| OpenVLM MMBench V1.1 EN - OCR | 100 | Accuracy (%) | 98.2 |
| OpenVLM MMBench V1.1 EN - Attribute Recognition | 97.8 | Accuracy (%) | 98.1 |
| OpenVLM AI2D - Moon Phase and Equinox | 79.8 | Accuracy (%) | 97.3 |
| OpenVLM MathVista - MWP | 87.1 | Accuracy (%) | 95.4 |
| OpenVLM MMStar - Fine-Grained Perception | 61.6 | Accuracy (%) | 94.7 |
| OpenVLM MMStar - Logical Reasoning | 77.6 | Accuracy (%) | 94.5 |
| OpenVLM AI2D - Solar System | 97.2 | Accuracy (%) | 94.3 |
| OpenVLM AI2D - Water, Carbon, and Nitrogen Cycle | 77.3 | Accuracy (%) | 94 |
Interactive version: theaggregate.ai/model?slug=sail-vl-1-5-8b · How It Works · Data refreshed daily, snapshot 2026-11-08.