# 4D Object Captioning on Objaverse 4,000 IDs (test)

Source: [4D Object Captioning on Objaverse 4,000 IDs (test)](<https://www.sota2.com/research/sota/4d-object-captioning-on-objaverse-4-000-ids-test>)

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

Data updated: 2026-02-26T05:45:46.544Z

Evaluation of 4D Object Captioning models on the specific 4,000 ID subset of the Objaverse test set, utilizing metrics like GPT-4, S-BERT, SimCSE, BLEU-1, ROUGE-L, and METEOR.

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 | 4D Object Captioning |
| Dataset (including any stated split) | Objaverse 4,000 IDs (test) |
| Metric | GPT-4 Score |
| Unit | PERCENT |
| Direction | Higher is better |
| Results loaded | 6 of 6 |
| 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

- [GPT-4 Score](<https://www.sota2.com/research/sota/4d-object-captioning-on-objaverse-4-000-ids-test.md?metric=GPT-4%20Score>): Higher is better; unit: PERCENT.

- [S-BERT Score](<https://www.sota2.com/research/sota/4d-object-captioning-on-objaverse-4-000-ids-test.md?metric=S-BERT%20Score>): Higher is better; unit: PERCENT.

- [SimCSE Score](<https://www.sota2.com/research/sota/4d-object-captioning-on-objaverse-4-000-ids-test.md?metric=SimCSE%20Score>): Higher is better; unit: PERCENT.

- [BLEU-1](<https://www.sota2.com/research/sota/4d-object-captioning-on-objaverse-4-000-ids-test.md?metric=BLEU-1>): Higher is better; unit: PERCENT.

- [ROUGE-L](<https://www.sota2.com/research/sota/4d-object-captioning-on-objaverse-4-000-ids-test.md?metric=ROUGE-L>): Higher is better; unit: PERCENT.

- [METEOR](<https://www.sota2.com/research/sota/4d-object-captioning-on-objaverse-4-000-ids-test.md?metric=METEOR>): Higher is better; unit: PERCENT.

## Recorded results

| Method | GPT-4 Score | Result date | Report paper |
| --- | --- | --- | --- |
| 4DPC^2hat | 73.27 | Not recorded | 698423f3a8b01e76cc8ee638 |
| MiniGPT-3D | 54.7 | Not recorded | 698423f3a8b01e76cc8ee638 |
| ShapeLLM-13B | 53.43 | Not recorded | 698423f3a8b01e76cc8ee638 |
| ShapeLLM-7B | 50.43 | Not recorded | 698423f3a8b01e76cc8ee638 |
| PointLLM-13B | 49.53 | Not recorded | 698423f3a8b01e76cc8ee638 |
| PointLLM-7B | 47.98 | Not recorded | 698423f3a8b01e76cc8ee638 |

### 4DPC^2hat

GPT-4 Score: 73.27. Result ID: 6997522e5698665ec4a3fd4e.

[Reporting paper](<https://www.sota2.com/research/paper/698423f3a8b01e76cc8ee638.md>)

[Method origin paper](<https://www.sota2.com/research/paper/698423f3a8b01e76cc8ee638.md>)

Recorded evaluation configuration: {"Input":"3D Point Cloud Sequence"}. This record does not assert independent reproduction.

### MiniGPT-3D

GPT-4 Score: 54.7. Result ID: 6997522e5698665ec4a3fd4d.

[Reporting paper](<https://www.sota2.com/research/paper/698423f3a8b01e76cc8ee638.md>)

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

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

Recorded evaluation configuration: {"Input":"3D+TA"}. This record does not assert independent reproduction.

### ShapeLLM-13B

GPT-4 Score: 53.43. Result ID: 6997522e5698665ec4a3fd4c.

[Reporting paper](<https://www.sota2.com/research/paper/698423f3a8b01e76cc8ee638.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa740c1e209131cea5f73a.md>)

[Implementation](<https://qizekun.github.io/ShapeLLM>)

Recorded evaluation configuration: {"Input":"3D+TA"}. This record does not assert independent reproduction.

### ShapeLLM-7B

GPT-4 Score: 50.43. Result ID: 6997522e5698665ec4a3fd4b.

[Reporting paper](<https://www.sota2.com/research/paper/698423f3a8b01e76cc8ee638.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa740c1e209131cea5f73a.md>)

[Implementation](<https://qizekun.github.io/ShapeLLM>)

Recorded evaluation configuration: {"Input":"3D+TA"}. This record does not assert independent reproduction.

### PointLLM-13B

GPT-4 Score: 49.53. Result ID: 6997522e5698665ec4a3fd4a.

[Reporting paper](<https://www.sota2.com/research/paper/698423f3a8b01e76cc8ee638.md>)

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

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

Recorded evaluation configuration: {"Input":"3D+TA"}. This record does not assert independent reproduction.

### PointLLM-7B

GPT-4 Score: 47.98. Result ID: 6997522e5698665ec4a3fd49.

[Reporting paper](<https://www.sota2.com/research/paper/698423f3a8b01e76cc8ee638.md>)

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

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

Recorded evaluation configuration: {"Input":"3D+TA"}. This record does not assert independent reproduction.

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