# Relative Position Reasoning on Scene 3 Laboratory storage 1.0 (overall) Source: [Relative Position Reasoning on Scene 3 Laboratory storage 1.0 (overall)]() Collection index: [llms.txt]() Data updated: 2026-02-26T05:26:59.413Z Evaluation of model performance on Relative Position Reasoning tasks within the Scene 3 Laboratory storage 1.0 dataset, measured by Accrel scores. A research result applies to its stated dataset, split, metric and evaluation configuration. Compare like-for-like settings, preserve the metric's unit and direction, and distinguish paper-reported results from independent reproduction. The best result recorded here is not a claim of exhaustive, current global coverage. | Scope | Recorded value | | --- | --- | | Task | Relative Position Reasoning | | Dataset (including any stated split) | Scene 3 Laboratory storage 1.0 (overall) | | Metric | Accrel | | Unit | PERCENT | | Direction | Higher is better | | Results loaded | 8 of 8 | | Coverage | All results currently returned for this benchmark | Evaluation settings may differ between results. The original papers and recorded configurations are needed to establish comparability. A shared metric name alone does not establish an identical protocol. ## Metrics - [Accrel](): Higher is better; unit: PERCENT. - [Accrel (Wall)](): Higher is better; unit: PERCENT. - [Accrel (Door)](): Higher is better; unit: PERCENT. - [Accrel (Window)](): Higher is better; unit: PERCENT. ## Recorded results | Method | Accrel | Result date | Report paper | | --- | --- | --- | --- | | gemini | 74 | Not recorded | 6971af2de25ad2fa9d1edabf | | SpatialMem | 74 | Not recorded | 6971af2de25ad2fa9d1edabf | | internvl (Entry 1) | 72 | Not recorded | 6971af2de25ad2fa9d1edabf | | llava | 70 | Not recorded | 6971af2de25ad2fa9d1edabf | | internvl (Entry 2) | 66 | Not recorded | 6971af2de25ad2fa9d1edabf | | openai | 64 | Not recorded | 6971af2de25ad2fa9d1edabf | | together | 58 | Not recorded | 6971af2de25ad2fa9d1edabf | | qwen25vl | 50 | Not recorded | 6971af2de25ad2fa9d1edabf | ### gemini Accrel: 74. Result ID: 69979f325698665ec4f52f0f. [Reporting paper]() Recorded evaluation configuration: {"Model":"Google Gemini"}. This record does not assert independent reproduction. ### SpatialMem Accrel: 74. Result ID: 69979f325698665ec4f52f16. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: Not recorded. This record does not assert independent reproduction. ### internvl (Entry 1) Accrel: 72. Result ID: 69979f325698665ec4f52f10. [Reporting paper]() Recorded evaluation configuration: {"Model":"InternVL 2.5","Deployment":"Local"}. This record does not assert independent reproduction. ### llava Accrel: 70. Result ID: 69979f325698665ec4f52f11. [Reporting paper]() Recorded evaluation configuration: {"Model":"LLaVA-Video","Deployment":"Local"}. This record does not assert independent reproduction. ### internvl (Entry 2) Accrel: 66. Result ID: 69979f325698665ec4f52f12. [Reporting paper]() Recorded evaluation configuration: {"Model":"InternVL 2.5","Deployment":"Local"}. This record does not assert independent reproduction. ### openai Accrel: 64. Result ID: 69979f325698665ec4f52f13. [Reporting paper]() Recorded evaluation configuration: {"Model":"OpenAI GPT-4o"}. This record does not assert independent reproduction. ### together Accrel: 58. Result ID: 69979f325698665ec4f52f14. [Reporting paper]() Recorded evaluation configuration: {"Model":"Qwen2.5-VL","Deployment":"Cloud"}. This record does not assert independent reproduction. ### qwen25vl Accrel: 50. Result ID: 69979f325698665ec4f52f15. [Reporting paper]() Recorded evaluation configuration: {"Model":"Qwen2.5-VL","Deployment":"Local"}. This record does not assert independent reproduction. Page 1 of 1. This is an index page, not the full collection.