Janus-Pro-7B: benchmark results
Provider: DeepSeek. Released 2025-01-27. Access: Open.
Aggregate Rating 1361 ± 16, rank #7415 of 10916 rated models, from 633 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| Med-R2 - Image Quality | 74.6 | Accuracy (%) on the Med-R2 Image Quality task (judging the q | 100 |
| OpenVLM MMT-Bench - Image Based Action Recognition | 100 | Score (%) | 100 |
| OpenVLM MMT-Bench - Plant Recognition | 95 | Score (%) | 98.1 |
| OpenVLM A-Bench - Nouns as Adjectives | 86 | Accuracy (%) | 97.2 |
| OpenVLM MMT-Bench - Small Object Detection | 70 | Score (%) | 96.8 |
| MedQ-Deg - Imaging Perception | 37 | Accuracy (%) on the Imaging Perception questions of MedQ-Deg | 94.9 |
| OpenVLM MMT-Bench - Spot The Diff | 85 | Score (%) | 92.7 |
| OpenVLM MMT-Bench - Attribute Hallucination | 85 | Score (%) | 92.5 |
| OpenVLM MMT-Bench - Painting Recognition | 90 | Score (%) | 90.8 |
| OpenVLM MMT-Bench - Image Matting | 30 | Score (%) | 90.3 |
| Open LMM Reasoning - WeMath - InsufficientKnowledge (Loose) | 65.7 | Accuracy (%) | 90.1 |
| OpenVLM MMT-Bench - Counting By Visual Prompting | 45 | Score (%) | 89.1 |
Interactive version: theaggregate.ai/model?slug=janus-pro-7b · How It Works · Data refreshed daily, snapshot 2026-11-08.