# Zero-shot understanding of embodied scenes on OpenEQA

Source: [Zero-shot understanding of embodied scenes on OpenEQA](<https://www.sota2.com/research/sota/zero-shot-understanding-of-embodied-scenes-on-openeqa>)

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

Data updated: 2026-02-26T06:03:55.592Z

Evaluation of model performance on the OpenEQA dataset focusing on zero-shot understanding capabilities within embodied scenes.

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 | Zero-shot understanding of embodied scenes |
| Dataset (including any stated split) | OpenEQA |
| Metric | Score |
| 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

- [Score](<https://www.sota2.com/research/sota/zero-shot-understanding-of-embodied-scenes-on-openeqa.md?metric=Score>): Higher is better; unit: PERCENT.

## Recorded results

| Method | Score | Result date | Report paper |
| --- | --- | --- | --- |
| Fast-ThinkAct-3B | 51.2 | Not recorded | 69687483e25ad2fa9d1ec538 |
| RoboBrain2.0-3B | 50.1 | Not recorded | 69687483e25ad2fa9d1ec538 |
| GPT-4V | 49.6 | Not recorded | 69687483e25ad2fa9d1ec538 |
| Magma-8B | 49.1 | Not recorded | 69687483e25ad2fa9d1ec538 |
| ThinkAct-3B | 48.9 | Not recorded | 69687483e25ad2fa9d1ec538 |
| InternVL3-2B | 48.8 | Not recorded | 69687483e25ad2fa9d1ec538 |
| InternVL2.5-2B | 47.1 | Not recorded | 69687483e25ad2fa9d1ec538 |
| NVILA-2B | 47 | Not recorded | 69687483e25ad2fa9d1ec538 |
| Gemini-2.5-Flash | 45.3 | Not recorded | 69687483e25ad2fa9d1ec538 |
| Qwen2.5-VL-3B | 43.4 | Not recorded | 69687483e25ad2fa9d1ec538 |

### Fast-ThinkAct-3B

Score: 51.2. Result ID: 6997bc635698665ec412f92b.

[Reporting paper](<https://www.sota2.com/research/paper/69687483e25ad2fa9d1ec538.md>)

[Method origin paper](<https://www.sota2.com/research/paper/69687483e25ad2fa9d1ec538.md>)

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

### RoboBrain2.0-3B

Score: 50.1. Result ID: 6997bc635698665ec412f929.

[Reporting paper](<https://www.sota2.com/research/paper/69687483e25ad2fa9d1ec538.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa7b4d1e209131cea87a54.md>)

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

### GPT-4V

Score: 49.6. Result ID: 6997bc635698665ec412f922.

[Reporting paper](<https://www.sota2.com/research/paper/69687483e25ad2fa9d1ec538.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa727a1e209131cea56a1e.md>)

[Implementation](<https://github.com/openai/evals>)

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

### Magma-8B

Score: 49.1. Result ID: 6997bc635698665ec412f928.

[Reporting paper](<https://www.sota2.com/research/paper/69687483e25ad2fa9d1ec538.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa770b1e209131cea70b91.md>)

[Implementation](<https://microsoft.github.io/Magma>)

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

### ThinkAct-3B

Score: 48.9. Result ID: 6997bc635698665ec412f92a.

[Reporting paper](<https://www.sota2.com/research/paper/69687483e25ad2fa9d1ec538.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa7b601e209131cea87ffb.md>)

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

### InternVL3-2B

Score: 48.8. Result ID: 6997bc635698665ec412f925.

[Reporting paper](<https://www.sota2.com/research/paper/69687483e25ad2fa9d1ec538.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa78241e209131cea76fa5.md>)

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

### InternVL2.5-2B

Score: 47.1. Result ID: 6997bc635698665ec412f924.

[Reporting paper](<https://www.sota2.com/research/paper/69687483e25ad2fa9d1ec538.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa7bab1e209131cea89c7f.md>)

[Implementation](<https://github.com/OpenGVLab/InternVL>)

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

### NVILA-2B

Score: 47. Result ID: 6997bc635698665ec412f926.

[Reporting paper](<https://www.sota2.com/research/paper/69687483e25ad2fa9d1ec538.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa77481e209131cea723b2.md>)

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

### Gemini-2.5-Flash

Score: 45.3. Result ID: 6997bc635698665ec412f923.

[Reporting paper](<https://www.sota2.com/research/paper/69687483e25ad2fa9d1ec538.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa7c311e209131cea8cef5.md>)

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

### Qwen2.5-VL-3B

Score: 43.4. Result ID: 6997bc635698665ec412f927.

[Reporting paper](<https://www.sota2.com/research/paper/69687483e25ad2fa9d1ec538.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa773a1e209131cea71d3c.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.
