# Indoor 3D Object Grounding on ARKitScenes (test)

Source: [Indoor 3D Object Grounding on ARKitScenes (test)](<https://www.sota2.com/research/sota/indoor-3d-object-grounding-on-arkitscenes-test>)

Collection index: [llms.txt](<https://www.sota2.com/research/llms.txt>)

Data updated: 2026-02-26T05:23:02.418Z

Evaluation of Indoor 3D Object Grounding performance on the test split of the ARKitScenes dataset, measured using category AP and category+location AP metrics.

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 | Indoor 3D Object Grounding |
| Dataset (including any stated split) | ARKitScenes (test) |
| Metric | AP (3D Cat) |
| Unit | PERCENT |
| Direction | Higher is better |
| Results loaded | 3 of 3 |
| 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

- [AP (3D Cat)](<https://www.sota2.com/research/sota/indoor-3d-object-grounding-on-arkitscenes-test.md?metric=AP%20(3D%20Cat)>): Higher is better; unit: PERCENT.

- [AP (3D Cat+Loc)](<https://www.sota2.com/research/sota/indoor-3d-object-grounding-on-arkitscenes-test.md?metric=AP%20(3D%20Cat%2BLoc)>): Higher is better; unit: PERCENT.

## Recorded results

| Method | AP (3D Cat) | Result date | Report paper |
| --- | --- | --- | --- |
| LocateAnything3D | 41.7 | Not recorded | 692689c7d6b0dd3342930655 |
| Cube-LLM\_large | 23.5 | Not recorded | 692689c7d6b0dd3342930655 |
| Cube-LLM\_small | 21.6 | Not recorded | 692689c7d6b0dd3342930655 |

### LocateAnything3D

AP (3D Cat): 41.7. Result ID: 699d4e055698665ec4f055ad.

[Reporting paper](<https://www.sota2.com/research/paper/692689c7d6b0dd3342930655.md>)

[Method origin paper](<https://www.sota2.com/research/paper/692689c7d6b0dd3342930655.md>)

Recorded evaluation configuration: {"Training Data":"curated dataset (1.7M images)"}. This record does not assert independent reproduction.

### Cube-LLM\_large

AP (3D Cat): 23.5. Result ID: 699d4e055698665ec4f055ac.

[Reporting paper](<https://www.sota2.com/research/paper/692689c7d6b0dd3342930655.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa73251e209131cea5aab3.md>)

Recorded evaluation configuration: {"Training Data":"full LV3D dataset (9.6M images)"}. This record does not assert independent reproduction.

### Cube-LLM\_small

AP (3D Cat): 21.6. Result ID: 699d4e055698665ec4f055ab.

[Reporting paper](<https://www.sota2.com/research/paper/692689c7d6b0dd3342930655.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa73251e209131cea5aab3.md>)

Recorded evaluation configuration: {"Training Data":"LV3D-small subset"}. This record does not assert independent reproduction.

Page 1 of 1. This is an index page, not the full collection.
