# 3D Video Object Detection on EmbodiedScan v1.0 (test)

Source: [3D Video Object Detection on EmbodiedScan v1.0 (test)](<https://www.sota2.com/research/sota/3d-video-object-detection-on-embodiedscan-v1-0-test>)

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

Data updated: 2026-03-18T08:51:59.606Z

Evaluation of 3D Video Object Detection performance on the test split of the EmbodiedScan v1.0 dataset, reporting P25, R25, and F125 metrics (likely based on IoU=0.25).

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 | 3D Video Object Detection |
| Dataset (including any stated split) | EmbodiedScan v1.0 (test) |
| Metric | P@0.25 |
| Unit | PERCENT |
| Direction | Higher is better |
| Results loaded | 10 of 10 |
| 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

- [P@0.25](<https://www.sota2.com/research/sota/3d-video-object-detection-on-embodiedscan-v1-0-test.md?metric=P%400.25>): Higher is better; unit: PERCENT.

- [R@0.25](<https://www.sota2.com/research/sota/3d-video-object-detection-on-embodiedscan-v1-0-test.md?metric=R%400.25>): Higher is better; unit: PERCENT.

- [F1@0.25](<https://www.sota2.com/research/sota/3d-video-object-detection-on-embodiedscan-v1-0-test.md?metric=F1%400.25>): Higher is better; unit: PERCENT.

## Recorded results

| Method | P@0.25 | Result date | Report paper |
| --- | --- | --- | --- |
| GAP-MLLM-3B | 54.2 | Not recorded | 69ba318b0636e5bf4afc0804 |
| GAP-MLLM-3B | 52.9 | Not recorded | 69ba318b0636e5bf4afc0804 |
| VG-LLM-4B (w/ GAP) | 48.8 | Not recorded | 69ba318b0636e5bf4afc0804 |
| VG-LLM-4B (w/ GAP) | 47 | Not recorded | 69ba318b0636e5bf4afc0804 |
| VG-LLM-7B | 43.5 | Not recorded | 69ba318b0636e5bf4afc0804 |
| VG-LLM-7B | 43.4 | Not recorded | 69ba318b0636e5bf4afc0804 |
| SpatialLM | 42.1 | Not recorded | 69ba318b0636e5bf4afc0804 |
| VG-LLM-4B | 41.7 | Not recorded | 69ba318b0636e5bf4afc0804 |
| SpatialLM | 40.8 | Not recorded | 69ba318b0636e5bf4afc0804 |
| VG-LLM-4B | 39.7 | Not recorded | 69ba318b0636e5bf4afc0804 |

### GAP-MLLM-3B

P@0.25: 54.2. Result ID: 69ba670247f34ce924a46ab5.

[Reporting paper](<https://www.sota2.com/research/paper/69ba318b0636e5bf4afc0804.md>)

[Method origin paper](<https://www.sota2.com/research/paper/69ba318b0636e5bf4afc0804.md>)

Recorded evaluation configuration: {"3D Input":"false","Frame Setting":"4-frame"}. This record does not assert independent reproduction.

### GAP-MLLM-3B

P@0.25: 52.9. Result ID: 69ba670247f34ce924a46ab6.

[Reporting paper](<https://www.sota2.com/research/paper/69ba318b0636e5bf4afc0804.md>)

[Method origin paper](<https://www.sota2.com/research/paper/69ba318b0636e5bf4afc0804.md>)

Recorded evaluation configuration: {"3D Input":"false","Frame Setting":"6-frame"}. This record does not assert independent reproduction.

### VG-LLM-4B (w/ GAP)

P@0.25: 48.8. Result ID: 69ba670247f34ce924a46ab4.

[Reporting paper](<https://www.sota2.com/research/paper/69ba318b0636e5bf4afc0804.md>)

[Method origin paper](<https://www.sota2.com/research/paper/69ba318b0636e5bf4afc0804.md>)

Recorded evaluation configuration: {"3D Input":"false","Frame Setting":"6-frame"}. This record does not assert independent reproduction.

### VG-LLM-4B (w/ GAP)

P@0.25: 47. Result ID: 69ba670247f34ce924a46ab3.

[Reporting paper](<https://www.sota2.com/research/paper/69ba318b0636e5bf4afc0804.md>)

[Method origin paper](<https://www.sota2.com/research/paper/69ba318b0636e5bf4afc0804.md>)

Recorded evaluation configuration: {"3D Input":"false","Frame Setting":"4-frame"}. This record does not assert independent reproduction.

### VG-LLM-7B

P@0.25: 43.5. Result ID: 69ba670247f34ce924a46ab2.

[Reporting paper](<https://www.sota2.com/research/paper/69ba318b0636e5bf4afc0804.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa7c701e209131cea8e12b.md>)

Recorded evaluation configuration: {"3D Input":"false","Frame Setting":"6-frame"}. This record does not assert independent reproduction.

### VG-LLM-7B

P@0.25: 43.4. Result ID: 69ba670247f34ce924a46ab1.

[Reporting paper](<https://www.sota2.com/research/paper/69ba318b0636e5bf4afc0804.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa7c701e209131cea8e12b.md>)

Recorded evaluation configuration: {"3D Input":"false","Frame Setting":"4-frame"}. This record does not assert independent reproduction.

### SpatialLM

P@0.25: 42.1. Result ID: 69ba670247f34ce924a46aae.

[Reporting paper](<https://www.sota2.com/research/paper/69ba318b0636e5bf4afc0804.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa79451e209131cea7cfad.md>)

Recorded evaluation configuration: {"3D Input":"true","Frame Setting":"6-frame"}. This record does not assert independent reproduction.

### VG-LLM-4B

P@0.25: 41.7. Result ID: 69ba670247f34ce924a46aaf.

[Reporting paper](<https://www.sota2.com/research/paper/69ba318b0636e5bf4afc0804.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa7c701e209131cea8e12b.md>)

Recorded evaluation configuration: {"3D Input":"false","Frame Setting":"4-frame"}. This record does not assert independent reproduction.

### SpatialLM

P@0.25: 40.8. Result ID: 69ba670247f34ce924a46aad.

[Reporting paper](<https://www.sota2.com/research/paper/69ba318b0636e5bf4afc0804.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa79451e209131cea7cfad.md>)

Recorded evaluation configuration: {"3D Input":"true","Frame Setting":"4-frame"}. This record does not assert independent reproduction.

### VG-LLM-4B

P@0.25: 39.7. Result ID: 69ba670247f34ce924a46ab0.

[Reporting paper](<https://www.sota2.com/research/paper/69ba318b0636e5bf4afc0804.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa7c701e209131cea8e12b.md>)

Recorded evaluation configuration: {"3D Input":"false","Frame Setting":"6-frame"}. This record does not assert independent reproduction.

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