OmniCoT - Locate: leaderboard
Locate questions of OmniCoT (2026), a panoramic spatial reasoning benchmark: accuracy in percent on multi-hop object and direction identification over 360-degree images, with chain-of-thought prompting. Higher is better.
Metric: Accuracy (%) on Locate questions (multi-hop object and direction identification) with chain-of-thought prompting on OmniCoT-B (synthetic 360-degree equirectangular panoramas with three anchor objects given by coordinates and orientation); answers are object identities, visibility, directions or the closest angle option, exact match after normalization. Source: arxiv.org. Saturation forecast: Not forecast. 21 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | GPT-5 | 24.4 | #167 |
| 2 | Seed 1.8 | 23.87 | #242 |
| 3 | Qwen 2.5 VL 72B Instruct | 23.69 | #846 |
| 4 | Qwen 2.5 VL 32B Instruct | 21.93 | #1053 |
| 5 | Grok 4 | 19.3 | #307 |
| 6 | Qwen 3 VL 8B Instruct | 18.73 | #981 |
| 7 | Qwen 2.5 VL 7B Instruct | 15.53 | #1833 |
| 8 | ChatGPT 4o | 15.18 | #610 |
| 9 | GLM-4.5V | 15.05 | #744 |
| 10 | Gemini 3 Flash | 14.17 | #164 |
No result here: #1 Claude Fable 5.2, #2 Gemini 4 Carbon, #5 Claude Opus 5.5.
Interactive version: theaggregate.ai/benchmark?slug=omnicot-locate · How It Works · Data refreshed daily, snapshot 2026-11-13.