VITA-1.5: benchmark results
Provider: Tencent. Released 2025-01-06. Access: Open.
Aggregate Rating 1465 ± 15, rank #3049 of 10916 rated models, from 610 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| OpenVLM MME - Scene | 178.8 | Score | 100 |
| OpenVLM MME - Celebrity | 181.5 | Score | 99.1 |
| OpenVLM MMBench V1.1 CN - OCR | 100 | Accuracy (%) | 98.6 |
| OpenVLM MMBench V1.1 EN - OCR | 100 | Accuracy (%) | 98.2 |
| OpenVLM MME - Landmark | 186.5 | Score | 98.1 |
| OpenVLM MMT-Bench - Disaster Recognition | 100 | Score (%) | 96.8 |
| OpenVLM MMT-Bench - Geometrical Perspective | 91.7 | Score (%) | 96.6 |
| OpenVLM MMT-Bench - Single Object Tracking | 75 | Score (%) | 96.6 |
| OpenVLM MMT-Bench - Furniture Keypoint Detection | 70 | Score (%) | 95.6 |
| OpenVLM SEEDBench IMG - Visual Reasoning | 82.2 | Accuracy (%) | 93.6 |
| OpenVLM MME - Perception | 1722.7 | Score | 92.7 |
| OpenVLM MMT-Bench - Attribute Hallucination | 85 | Score (%) | 92.5 |
Interactive version: theaggregate.ai/model?slug=vita-1-5 · How It Works · Data refreshed daily, snapshot 2026-11-08.