# Spatial Understanding on CV-Bench 2D Overall

Source: [Spatial Understanding on CV-Bench 2D Overall](<https://www.sota2.com/research/sota/spatial-understanding-on-cv-bench-2d-overall>)

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

Data updated: 2026-02-26T05:31:36.491Z

Evaluation of model performance on the CV-Bench 2D Overall dataset focusing on Spatial Understanding capabilities, measured by Accuracy.

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 | Spatial Understanding |
| Dataset (including any stated split) | CV-Bench 2D Overall |
| Metric | Accuracy |
| Unit | PERCENT |
| Direction | Higher is better |
| Results loaded | 15 of 15 |
| 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

- [Accuracy](<https://www.sota2.com/research/sota/spatial-understanding-on-cv-bench-2d-overall.md?metric=Accuracy>): Higher is better; unit: PERCENT.

## Recorded results

| Method | Accuracy | Result date | Report paper |
| --- | --- | --- | --- |
| Qwen2.5-VL | 75.4 | Not recorded | 6969c5dde25ad2fa9d1ec686 |
| LLaVA-OneVision | 73 | Not recorded | 6969c5dde25ad2fa9d1ec686 |
| Molmo | 72.8 | Not recorded | 6969c5dde25ad2fa9d1ec686 |
| Gemma3-4b-it | 65.9 | Not recorded | 6969c5dde25ad2fa9d1ec686 |
| PaliGemma | 62.4 | Not recorded | 6969c5dde25ad2fa9d1ec686 |
| LLaVA-NeXT | 60.6 | Not recorded | 6969c5dde25ad2fa9d1ec686 |
| SmolVLM2 | 57.7 | Not recorded | 6969c5dde25ad2fa9d1ec686 |
| LLaVA v1.5 | 49 | Not recorded | 6969c5dde25ad2fa9d1ec686 |
| LLaVA-AIMv2 | 46.6 | Not recorded | 6969c5dde25ad2fa9d1ec686 |
| LLaVA-2D-RoPE | 44.3 | Not recorded | 6969c5dde25ad2fa9d1ec686 |
| LLaVA-SigLIP2 | 44.2 | Not recorded | 6969c5dde25ad2fa9d1ec686 |
| LLaVA-AIMv2-2D-RoPE | 43.2 | Not recorded | 6969c5dde25ad2fa9d1ec686 |
| LLaVA-SigLIP-2D-RoPE | 42.5 | Not recorded | 6969c5dde25ad2fa9d1ec686 |
| LLaVA-SigLIP2-2D-RoPE | 41.5 | Not recorded | 6969c5dde25ad2fa9d1ec686 |
| LLaVA-SigLIP | 41.2 | Not recorded | 6969c5dde25ad2fa9d1ec686 |

### Qwen2.5-VL

Accuracy: 75.4. Result ID: 6997baba5698665ec4114486.

[Reporting paper](<https://www.sota2.com/research/paper/6969c5dde25ad2fa9d1ec686.md>)

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

Recorded evaluation configuration: Not recorded. This record does not assert independent reproduction.

### LLaVA-OneVision

Accuracy: 73. Result ID: 6997baba5698665ec4114485.

[Reporting paper](<https://www.sota2.com/research/paper/6969c5dde25ad2fa9d1ec686.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa756f1e209131cea67e58.md>)

[Implementation](<https://llava-vl.github.io/blog/llava-onevision>)

Recorded evaluation configuration: Not recorded. This record does not assert independent reproduction.

### Molmo

Accuracy: 72.8. Result ID: 6997baba5698665ec411448a.

[Reporting paper](<https://www.sota2.com/research/paper/6969c5dde25ad2fa9d1ec686.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa76101e209131cea6b37c.md>)

[Implementation](<https://molmo.allenai.org/>)

Recorded evaluation configuration: Not recorded. This record does not assert independent reproduction.

### Gemma3-4b-it

Accuracy: 65.9. Result ID: 6997baba5698665ec4114488.

[Reporting paper](<https://www.sota2.com/research/paper/6969c5dde25ad2fa9d1ec686.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa779f1e209131cea746cd.md>)

Recorded evaluation configuration: Not recorded. This record does not assert independent reproduction.

### PaliGemma

Accuracy: 62.4. Result ID: 6997baba5698665ec4114489.

[Reporting paper](<https://www.sota2.com/research/paper/6969c5dde25ad2fa9d1ec686.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa751e1e209131cea65e0c.md>)

Recorded evaluation configuration: Not recorded. This record does not assert independent reproduction.

### LLaVA-NeXT

Accuracy: 60.6. Result ID: 6997baba5698665ec4114484.

[Reporting paper](<https://www.sota2.com/research/paper/6969c5dde25ad2fa9d1ec686.md>)

Recorded evaluation configuration: Not recorded. This record does not assert independent reproduction.

### SmolVLM2

Accuracy: 57.7. Result ID: 6997baba5698665ec4114487.

[Reporting paper](<https://www.sota2.com/research/paper/6969c5dde25ad2fa9d1ec686.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa77d51e209131cea75788.md>)

Recorded evaluation configuration: Not recorded. This record does not assert independent reproduction.

### LLaVA v1.5

Accuracy: 49. Result ID: 6997baba5698665ec411448b.

[Reporting paper](<https://www.sota2.com/research/paper/6969c5dde25ad2fa9d1ec686.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa732e1e209131cea5af59.md>)

[Implementation](<https://llava-vl.github.io/>)

Recorded evaluation configuration: Not recorded. This record does not assert independent reproduction.

### LLaVA-AIMv2

Accuracy: 46.6. Result ID: 6997baba5698665ec4114491.

[Reporting paper](<https://www.sota2.com/research/paper/6969c5dde25ad2fa9d1ec686.md>)

[Method origin paper](<https://www.sota2.com/research/paper/6969c5dde25ad2fa9d1ec686.md>)

Recorded evaluation configuration: {"encoder":"AIMv2"}. This record does not assert independent reproduction.

### LLaVA-2D-RoPE

Accuracy: 44.3. Result ID: 6997baba5698665ec411448c.

[Reporting paper](<https://www.sota2.com/research/paper/6969c5dde25ad2fa9d1ec686.md>)

[Method origin paper](<https://www.sota2.com/research/paper/6969c5dde25ad2fa9d1ec686.md>)

Recorded evaluation configuration: {"positional\_embedding":"2D-RoPE"}. This record does not assert independent reproduction.

### LLaVA-SigLIP2

Accuracy: 44.2. Result ID: 6997baba5698665ec411448f.

[Reporting paper](<https://www.sota2.com/research/paper/6969c5dde25ad2fa9d1ec686.md>)

[Method origin paper](<https://www.sota2.com/research/paper/6969c5dde25ad2fa9d1ec686.md>)

Recorded evaluation configuration: {"encoder":"SigLIP2"}. This record does not assert independent reproduction.

### LLaVA-AIMv2-2D-RoPE

Accuracy: 43.2. Result ID: 6997baba5698665ec4114492.

[Reporting paper](<https://www.sota2.com/research/paper/6969c5dde25ad2fa9d1ec686.md>)

[Method origin paper](<https://www.sota2.com/research/paper/6969c5dde25ad2fa9d1ec686.md>)

Recorded evaluation configuration: {"encoder":"AIMv2","positional\_embedding":"2D-RoPE"}. This record does not assert independent reproduction.

### LLaVA-SigLIP-2D-RoPE

Accuracy: 42.5. Result ID: 6997baba5698665ec411448e.

[Reporting paper](<https://www.sota2.com/research/paper/6969c5dde25ad2fa9d1ec686.md>)

[Method origin paper](<https://www.sota2.com/research/paper/6969c5dde25ad2fa9d1ec686.md>)

Recorded evaluation configuration: {"encoder":"SigLIP","positional\_embedding":"2D-RoPE"}. This record does not assert independent reproduction.

### LLaVA-SigLIP2-2D-RoPE

Accuracy: 41.5. Result ID: 6997baba5698665ec4114490.

[Reporting paper](<https://www.sota2.com/research/paper/6969c5dde25ad2fa9d1ec686.md>)

[Method origin paper](<https://www.sota2.com/research/paper/6969c5dde25ad2fa9d1ec686.md>)

Recorded evaluation configuration: {"encoder":"SigLIP2","positional\_embedding":"2D-RoPE"}. This record does not assert independent reproduction.

### LLaVA-SigLIP

Accuracy: 41.2. Result ID: 6997baba5698665ec411448d.

[Reporting paper](<https://www.sota2.com/research/paper/6969c5dde25ad2fa9d1ec686.md>)

[Method origin paper](<https://www.sota2.com/research/paper/6969c5dde25ad2fa9d1ec686.md>)

Recorded evaluation configuration: {"encoder":"SigLIP"}. This record does not assert independent reproduction.

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