Spatiotemporal Intelligence on VSTI-Bench
77AccuracyHuman
Evaluation Results
| Method | Links | |
|---|---|---|
| HumanCategory=Baseline2026.02 | 77 | |
| VLM-3R-7BCategory=Open-Sourced spatial models2026.02 | 58.8 | |
| SSR-3DCategory=Open-Sourced spatial models2026.02 | 44.8 | |
| SSR-2DCategory=Open-Sourced spatial models2026.02 | 41.8 | |
| LLaVA-OneVision-7BCategory=Open-Sourced general models2026.02 | 41.7 | |
| Chance Level (Frequency)Category=Baseline2026.02 | 22.4 |