OST-Bench - Agent Object Spatial: leaderboard
The Agent Object Spatial questions of OST-Bench's multi-round version: the model judges the direction and distance of previously seen objects relative to its current position. Scores are accuracy (%); the README states 1.4k scenes and 10k question-answer pairs, the per-model sample is not stated.
Metric: Accuracy (%). Source: github.com. Saturation forecast: Estimated already saturated. 10 models tracked. Updated 17 Nov 2026.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | GPT-4.1 | 37.65 | #505 |
| 2 | LLaVA-Video-72B-Qwen2 | 35.07 | #1344 |
| 3 | InternVL2.5-8B | 34.61 | #2401 |
| 4 | InternVL2.5-38B | 33.95 | #1366 |
| 5 | Gemini 2.0 Flash | 33.8 | #723 |
| 6 | InternVL2.5-78B | 32.93 | #1036 |
| 7 | Qwen 2.5 VL 32B Instruct | 32.18 | #1074 |
| 8 | Qwen 2.5 VL 7B Instruct | 31.53 | #1893 |
| 9 | Qwen 2.5 VL 72B Instruct | 28.87 | #849 |
| 10 | LLaVA-Video-7B-Qwen2 | 28.8 | #1910 |
No result here: #1 Claude Fable 5.2, #2 GPT-6.2 Astra, #3 Gemini 4 Carbon.
Interactive version: theaggregate.ai/benchmark?slug=ost-bench-agent-object-spatial · How It Works · Data refreshed daily, snapshot 2026-11-17.