Hunyuan-T1-Vision: benchmark results
Provider: Tencent.
Aggregate Rating 1623 ± 31, rank #802 of 10916 rated models, from 21 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| FlagEval VQA - Text Understanding (Chinese) | 78.2 | Score | 87 |
| EPIC-Bench - Basic Attributes | 51.24 | Target localization score (0-100) on the basic-attribute tas | 86.4 |
| EPIC-Bench - Contact Relationship | 38.95 | Manipulation score (0-100) on contact relationship (localizi | 84.1 |
| EPIC-Bench - Feasible Path | 50.08 | Navigation path score (0-100) on feasible path planning (a p | 79.5 |
| EPIC-Bench | 43 | Average score (0-100) over all samples, each sample scored b | 78.4 |
| EPIC-Bench - Embodied Compositional Attributes | 39.67 | Target localization score (0-100) on the embodied compositio | 78.4 |
| EPIC-Bench - Spatial-Related Attributes | 47.6 | Target localization score (0-100) on the spatial-attribute t | 73.9 |
| EPIC-Bench - Ground Detection | 42.63 | Navigation score (0-100) on ground detection (localizing tra | 69.3 |
| EPIC-Bench - Affordance Region | 6.63 | Manipulation score (0-100) on affordance region localization | 58 |
| EPIC-Bench - Placement Region | 34.49 | Manipulation score (0-100) on placement region (where an obj | 53.4 |
| FlagEval VQA - Geographic Reasoning | 41.8 | Score | 47.8 |
| FlagEval VQA - Long-tail Recognition | 34.8 | Score | 47.8 |
Interactive version: theaggregate.ai/model?slug=hunyuan-t1-vision · How It Works · Data refreshed daily, snapshot 2026-11-08.