# 6DoF object manipulation trajectory generation on HOT3D

Source: [6DoF object manipulation trajectory generation on HOT3D](<https://www.sota2.com/research/sota/6dof-object-manipulation-trajectory-generation-on-hot3d>)

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

Data updated: 2026-04-03T07:21:33.158Z

Evaluation of models generating 6-Degree-of-Freedom object manipulation trajectories on the HOT3D dataset, typically assessed using positional ADE/FDE and GD 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 | 6DoF object manipulation trajectory generation |
| Dataset (including any stated split) | HOT3D |
| Metric | 3D Positional ADE |
| Unit | SCALAR |
| Direction | Lower is better |
| Results loaded | 19 of 19 |
| 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

- [3D Positional ADE](<https://www.sota2.com/research/sota/6dof-object-manipulation-trajectory-generation-on-hot3d.md?metric=3D%20Positional%20ADE>): Lower is better; unit: SCALAR.

- [3D Positional FDE](<https://www.sota2.com/research/sota/6dof-object-manipulation-trajectory-generation-on-hot3d.md?metric=3D%20Positional%20FDE>): Lower is better; unit: SCALAR.

- [2D Positional ADE](<https://www.sota2.com/research/sota/6dof-object-manipulation-trajectory-generation-on-hot3d.md?metric=2D%20Positional%20ADE>): Lower is better; unit: SCALAR.

- [2D Positional FDE](<https://www.sota2.com/research/sota/6dof-object-manipulation-trajectory-generation-on-hot3d.md?metric=2D%20Positional%20FDE>): Lower is better; unit: SCALAR.

- [Geometric Distance (GD)](<https://www.sota2.com/research/sota/6dof-object-manipulation-trajectory-generation-on-hot3d.md?metric=Geometric%20Distance%20(GD)>): Lower is better; unit: SCALAR.

## Recorded results

| Method | 3D Positional ADE | Result date | Report paper |
| --- | --- | --- | --- |
| EgoFlow | 0.265 | Not recorded | 69cf496c0636e5bf4afcd8e9 |
| PointLLM (7B) | 0.271 | Not recorded | 68fa79131e209131cea7bfd0 |
| BLIP-2 (2.7B) | 0.275 | Not recorded | 68fa79131e209131cea7bfd0 |
| BLIP-2 (2.7B) | 0.28 | Not recorded | 68fa79131e209131cea7bfd0 |
| MiniGPT-3D (2.7B) | 0.281 | Not recorded | 68fa79131e209131cea7bfd0 |
| BLIP-2 (6.7B) | 0.282 | Not recorded | 68fa79131e209131cea7bfd0 |
| BLIP-2 (6.7B) | 0.286 | Not recorded | 68fa79131e209131cea7bfd0 |
| VILA (8B) | 0.293 | Not recorded | 68fa79131e209131cea7bfd0 |
| VILA (3B) | 0.294 | Not recorded | 68fa79131e209131cea7bfd0 |
| GIMO | 0.299 | Not recorded | 69cf496c0636e5bf4afcd8e9 |
| VILA (8B) | 0.318 | Not recorded | 68fa79131e209131cea7bfd0 |
| Seq2Seq | 0.341 | Not recorded | 68fa79131e209131cea7bfd0 |
| Egoscaler | 0.351 | Not recorded | 69cf496c0636e5bf4afcd8e9 |
| Seq2Seq | 0.374 | Not recorded | 68fa79131e209131cea7bfd0 |
| VILA (3B) | 0.477 | Not recorded | 68fa79131e209131cea7bfd0 |
| CHOIS | 0.513 | Not recorded | 69cf496c0636e5bf4afcd8e9 |
| SPOT | 1.018 | Not recorded | 69cf496c0636e5bf4afcd8e9 |
| DP3 | 1.019 | Not recorded | 69cf496c0636e5bf4afcd8e9 |
| M2Diffuser | 1.079 | Not recorded | 69cf496c0636e5bf4afcd8e9 |

### EgoFlow

3D Positional ADE: 0.265. Result ID: 69cf6a6a47f34ce924b8db4f.

[Reporting paper](<https://www.sota2.com/research/paper/69cf496c0636e5bf4afcd8e9.md>)

[Method origin paper](<https://www.sota2.com/research/paper/69cf496c0636e5bf4afcd8e9.md>)

Recorded evaluation configuration: {"Zero-shot":"true"}. This record does not assert independent reproduction.

### PointLLM (7B)

3D Positional ADE: 0.271. Result ID: 6995c27f3badfaffe6c5472e.

[Reporting paper](<https://www.sota2.com/research/paper/68fa79131e209131cea7bfd0.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa74c51e209131cea63b44.md>)

[Implementation](<https://github.com/OpenRobotLab/PointLLM>)

Recorded evaluation configuration: {"Input":"Point cloud"}. This record does not assert independent reproduction.

### BLIP-2 (2.7B)

3D Positional ADE: 0.275. Result ID: 6995c27f3badfaffe6c54729.

[Reporting paper](<https://www.sota2.com/research/paper/68fa79131e209131cea7bfd0.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6fb71e209131cea46fe6.md>)

[Implementation](<https://github.com/salesforce/LAVIS/tree/main/projects/blip2>)

Recorded evaluation configuration: {"Input":"Image + Depth"}. This record does not assert independent reproduction.

### BLIP-2 (2.7B)

3D Positional ADE: 0.28. Result ID: 6995c27f3badfaffe6c54724.

[Reporting paper](<https://www.sota2.com/research/paper/68fa79131e209131cea7bfd0.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6fb71e209131cea46fe6.md>)

[Implementation](<https://github.com/salesforce/LAVIS/tree/main/projects/blip2>)

Recorded evaluation configuration: {"Input":"Image"}. This record does not assert independent reproduction.

### MiniGPT-3D (2.7B)

3D Positional ADE: 0.281. Result ID: 6995c27f3badfaffe6c5472d.

[Reporting paper](<https://www.sota2.com/research/paper/68fa79131e209131cea7bfd0.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa73161e209131cea5a71d.md>)

[Implementation](<https://github.com/TangYuan96/MiniGPT-3D>)

Recorded evaluation configuration: {"Input":"Point cloud"}. This record does not assert independent reproduction.

### BLIP-2 (6.7B)

3D Positional ADE: 0.282. Result ID: 6995c27f3badfaffe6c5472a.

[Reporting paper](<https://www.sota2.com/research/paper/68fa79131e209131cea7bfd0.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6fb71e209131cea46fe6.md>)

[Implementation](<https://github.com/salesforce/LAVIS/tree/main/projects/blip2>)

Recorded evaluation configuration: {"Input":"Image + Depth"}. This record does not assert independent reproduction.

### BLIP-2 (6.7B)

3D Positional ADE: 0.286. Result ID: 6995c27f3badfaffe6c54725.

[Reporting paper](<https://www.sota2.com/research/paper/68fa79131e209131cea7bfd0.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6fb71e209131cea46fe6.md>)

[Implementation](<https://github.com/salesforce/LAVIS/tree/main/projects/blip2>)

Recorded evaluation configuration: {"Input":"Image"}. This record does not assert independent reproduction.

### VILA (8B)

3D Positional ADE: 0.293. Result ID: 6995c27f3badfaffe6c54727.

[Reporting paper](<https://www.sota2.com/research/paper/68fa79131e209131cea7bfd0.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa73311e209131cea5b06b.md>)

[Implementation](<https://github.com/NVlabs/VILA>)

Recorded evaluation configuration: {"Input":"Image"}. This record does not assert independent reproduction.

### VILA (3B)

3D Positional ADE: 0.294. Result ID: 6995c27f3badfaffe6c5472b.

[Reporting paper](<https://www.sota2.com/research/paper/68fa79131e209131cea7bfd0.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa73311e209131cea5b06b.md>)

[Implementation](<https://github.com/NVlabs/VILA>)

Recorded evaluation configuration: {"Input":"Image + Depth"}. This record does not assert independent reproduction.

### GIMO

3D Positional ADE: 0.299. Result ID: 69cf6a6a47f34ce924b8db49.

[Reporting paper](<https://www.sota2.com/research/paper/69cf496c0636e5bf4afcd8e9.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6d171e209131cea379cf.md>)

Recorded evaluation configuration: {"Zero-shot":"true"}. This record does not assert independent reproduction.

### VILA (8B)

3D Positional ADE: 0.318. Result ID: 6995c27f3badfaffe6c5472c.

[Reporting paper](<https://www.sota2.com/research/paper/68fa79131e209131cea7bfd0.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa73311e209131cea5b06b.md>)

[Implementation](<https://github.com/NVlabs/VILA>)

Recorded evaluation configuration: {"Input":"Image + Depth"}. This record does not assert independent reproduction.

### Seq2Seq

3D Positional ADE: 0.341. Result ID: 6995c27f3badfaffe6c54728.

[Reporting paper](<https://www.sota2.com/research/paper/68fa79131e209131cea7bfd0.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa703d1e209131cea49db1.md>)

[Implementation](<https://github.com/tensorflow/tensor2tensor>)

Recorded evaluation configuration: {"Input":"Image + Depth"}. This record does not assert independent reproduction.

### Egoscaler

3D Positional ADE: 0.351. Result ID: 69cf6a6a47f34ce924b8db4e.

[Reporting paper](<https://www.sota2.com/research/paper/69cf496c0636e5bf4afcd8e9.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa79131e209131cea7bfd0.md>)

[Implementation](<https://biscue5.github.io/egoscaler-project-page/>)

Recorded evaluation configuration: {"Zero-shot":"true"}. This record does not assert independent reproduction.

### Seq2Seq

3D Positional ADE: 0.374. Result ID: 6995c27f3badfaffe6c54723.

[Reporting paper](<https://www.sota2.com/research/paper/68fa79131e209131cea7bfd0.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa703d1e209131cea49db1.md>)

[Implementation](<https://github.com/tensorflow/tensor2tensor>)

Recorded evaluation configuration: {"Input":"Image"}. This record does not assert independent reproduction.

### VILA (3B)

3D Positional ADE: 0.477. Result ID: 6995c27f3badfaffe6c54726.

[Reporting paper](<https://www.sota2.com/research/paper/68fa79131e209131cea7bfd0.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa73311e209131cea5b06b.md>)

[Implementation](<https://github.com/NVlabs/VILA>)

Recorded evaluation configuration: {"Input":"Image"}. This record does not assert independent reproduction.

### CHOIS

3D Positional ADE: 0.513. Result ID: 69cf6a6a47f34ce924b8db4a.

[Reporting paper](<https://www.sota2.com/research/paper/69cf496c0636e5bf4afcd8e9.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa74201e209131cea5fce5.md>)

Recorded evaluation configuration: {"Zero-shot":"true"}. This record does not assert independent reproduction.

### SPOT

3D Positional ADE: 1.018. Result ID: 69cf6a6a47f34ce924b8db4b.

[Reporting paper](<https://www.sota2.com/research/paper/69cf496c0636e5bf4afcd8e9.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa788c1e209131cea78ff3.md>)

Recorded evaluation configuration: {"Zero-shot":"true"}. This record does not assert independent reproduction.

### DP3

3D Positional ADE: 1.019. Result ID: 69cf6a6a47f34ce924b8db4c.

[Reporting paper](<https://www.sota2.com/research/paper/69cf496c0636e5bf4afcd8e9.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa74f31e209131cea64c76.md>)

[Implementation](<https://github.com/YanjieZe/3D-Diffusion-Policy>)

Recorded evaluation configuration: {"Zero-shot":"true"}. This record does not assert independent reproduction.

### M2Diffuser

3D Positional ADE: 1.079. Result ID: 69cf6a6a47f34ce924b8db4d.

[Reporting paper](<https://www.sota2.com/research/paper/69cf496c0636e5bf4afcd8e9.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa77cb1e209131cea755fd.md>)

Recorded evaluation configuration: {"Zero-shot":"true"}. This record does not assert independent reproduction.

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