DeepSeek-VL-1.3B: benchmark results
Provider: DeepSeek. Released 2024-03-11.
Aggregate Rating 1277 ± 38, rank #9001 of 10916 rated models, from 434 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| OpenVLM MMT-Bench - Gesture Recognition | 75 | Score (%) | 95.1 |
| OpenVLM MMT-Bench - Spot The Diff | 90 | Score (%) | 94.9 |
| OpenVLM MMT-Bench - MIA | 77.5 | Score (%) | 94.7 |
| OpenVLM MMT-Bench - Table Structure Recognition | 100 | Score (%) | 94.7 |
| OpenVLM MMT-Bench - Point Tracking | 85 | Score (%) | 93.9 |
| OpenVLM MMT-Bench - Scene Emotion Recognition | 65 | Score (%) | 93.2 |
| OpenVLM MMT-Bench - Transparent Object Detection | 80 | Score (%) | 92.7 |
| OpenVLM MMT-Bench - Micro Expression Recognition | 40 | Score (%) | 89.8 |
| OpenVLM MMT-Bench - Counting By Visual Prompting | 45 | Score (%) | 89.1 |
| OpenVLM MMT-Bench - Small Object Detection | 55 | Score (%) | 88.6 |
| OpenVLM MMT-Bench - Counting By Category | 38.8 | Score (%) | 85.4 |
| OpenVLM COCO Captions | 20.8 | CIDEr | 84.8 |
Interactive version: theaggregate.ai/model?slug=deepseek-vl-1-3b · How It Works · Data refreshed daily, snapshot 2026-11-08.