JT-VL-Chat-V3.0: benchmark results
Provider: China Mobile. Released 2025-07-26. Access: API.
Aggregate Rating 1662 ± 39, rank #541 of 10916 rated models, from 145 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| OpenVLM MMStar - Coarse Perception | 90.8 | Accuracy (%) | 100 |
| OpenVLM MMStar - Fine-Grained Perception | 80.4 | Accuracy (%) | 100 |
| OpenVLM MMStar - Instance Reasoning | 88.8 | Accuracy (%) | 100 |
| OpenVLM MathVista - NUM | 66 | Accuracy (%) | 100 |
| OpenVLM OCRBench | 95 | Score (normalized) | 100 |
| OpenVLM OCRBench - Scene Text-centric VQA | 99.5 | Correct Answers (%) | 100 |
| OpenVLM OCRBench - Text Recognition | 100 | Correct Answers (%) | 100 |
| OpenVLM MMBench V1.1 EN - Future Prediction | 89.3 | Accuracy (%) | 99.8 |
| OpenVLM MMBench V1.1 EN - Overall | 88.5 | Accuracy (%) | 99.8 |
| OpenVLM MMMU - Geography | 83.3 | Accuracy (%) | 99.8 |
| OpenVLM MathVista - VQA | 68.2 | Accuracy (%) | 99.8 |
| OpenVLM OCRBench - Handwritten Mathematical Expression Recognition | 92 | Correct Answers (%) | 99.8 |
Interactive version: theaggregate.ai/model?slug=jt-vl-chat-v3-0 · How It Works · Data refreshed daily, snapshot 2026-11-08.