# Spatial Geometric Constraints on MICON-Bench

Source: [Spatial Geometric Constraints on MICON-Bench](<https://www.sota2.com/research/sota/spatial-geometric-constraints-on-micon-bench>)

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

Data updated: 2026-02-26T05:38:31.910Z

Evaluation of model performance on the MICON-Bench dataset, specifically assessing adherence to Spatial Geometric Constraints using perceptual similarity metrics like CLIP Score, DINO v2, and LPIPS.

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 Geometric Constraints |
| Dataset (including any stated split) | MICON-Bench |
| Metric | CLIP Score (T2I) |
| Unit | SCALAR |
| Direction | Higher is better |
| Results loaded | 9 of 9 |
| 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

- [CLIP Score (T2I)](<https://www.sota2.com/research/sota/spatial-geometric-constraints-on-micon-bench.md?metric=CLIP%20Score%20(T2I)>): Higher is better; unit: SCALAR.

- [CLIP Score (I2I)](<https://www.sota2.com/research/sota/spatial-geometric-constraints-on-micon-bench.md?metric=CLIP%20Score%20(I2I)>): Higher is better; unit: SCALAR.

- [DINO v2 (I2I)](<https://www.sota2.com/research/sota/spatial-geometric-constraints-on-micon-bench.md?metric=DINO%20v2%20(I2I)>): Higher is better; unit: SCALAR.

- [LPIPS (I2I)](<https://www.sota2.com/research/sota/spatial-geometric-constraints-on-micon-bench.md?metric=LPIPS%20(I2I)>): Lower is better; unit: SCALAR.

## Recorded results

| Method | CLIP Score (T2I) | Result date | Report paper |
| --- | --- | --- | --- |
| OmniGen2 | 0.3828 | Not recorded | 699d30c46017c4460ffab7e4 |
| OmniGen2 | 0.3828 | Not recorded | 699d30c46017c4460ffab7e4 |
| BAGEL | 0.3761 | Not recorded | 699d30c46017c4460ffab7e4 |
| BAGEL | 0.3748 | Not recorded | 699d30c46017c4460ffab7e4 |
| GPT-Image | 0.3742 | Not recorded | 699d30c46017c4460ffab7e4 |
| Qwen-Image-Edit | 0.37 | Not recorded | 699d30c46017c4460ffab7e4 |
| Nano-Banana | 0.3665 | Not recorded | 699d30c46017c4460ffab7e4 |
| DreamOmni2 | 0.3647 | Not recorded | 699d30c46017c4460ffab7e4 |
| UNO | 0.3566 | Not recorded | 699d30c46017c4460ffab7e4 |

### OmniGen2

CLIP Score (T2I): 0.3828. Result ID: 699e4eb25698665ec40d7d70.

[Reporting paper](<https://www.sota2.com/research/paper/699d30c46017c4460ffab7e4.md>)

[Method origin paper](<https://www.sota2.com/research/paper/699d30c46017c4460ffab7e4.md>)

Recorded evaluation configuration: {"DAR":false}. This record does not assert independent reproduction.

### OmniGen2

CLIP Score (T2I): 0.3828. Result ID: 699e4eb25698665ec40d7d71.

[Reporting paper](<https://www.sota2.com/research/paper/699d30c46017c4460ffab7e4.md>)

[Method origin paper](<https://www.sota2.com/research/paper/699d30c46017c4460ffab7e4.md>)

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

### BAGEL

CLIP Score (T2I): 0.3761. Result ID: 699e4eb25698665ec40d7d6f.

[Reporting paper](<https://www.sota2.com/research/paper/699d30c46017c4460ffab7e4.md>)

[Method origin paper](<https://www.sota2.com/research/paper/699d30c46017c4460ffab7e4.md>)

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

### BAGEL

CLIP Score (T2I): 0.3748. Result ID: 699e4eb25698665ec40d7d6e.

[Reporting paper](<https://www.sota2.com/research/paper/699d30c46017c4460ffab7e4.md>)

[Method origin paper](<https://www.sota2.com/research/paper/699d30c46017c4460ffab7e4.md>)

Recorded evaluation configuration: {"DAR":false}. This record does not assert independent reproduction.

### GPT-Image

CLIP Score (T2I): 0.3742. Result ID: 699e4eb25698665ec40d7d6a.

[Reporting paper](<https://www.sota2.com/research/paper/699d30c46017c4460ffab7e4.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa75681e209131cea67a98.md>)

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

### Qwen-Image-Edit

CLIP Score (T2I): 0.37. Result ID: 699e4eb25698665ec40d7d6b.

[Reporting paper](<https://www.sota2.com/research/paper/699d30c46017c4460ffab7e4.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa7a6a1e209131cea83137.md>)

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

### Nano-Banana

CLIP Score (T2I): 0.3665. Result ID: 699e4eb25698665ec40d7d69.

[Reporting paper](<https://www.sota2.com/research/paper/699d30c46017c4460ffab7e4.md>)

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

### DreamOmni2

CLIP Score (T2I): 0.3647. Result ID: 699e4eb25698665ec40d7d6d.

[Reporting paper](<https://www.sota2.com/research/paper/699d30c46017c4460ffab7e4.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa7bec1e209131cea8b53f.md>)

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

### UNO

CLIP Score (T2I): 0.3566. Result ID: 699e4eb25698665ec40d7d6c.

[Reporting paper](<https://www.sota2.com/research/paper/699d30c46017c4460ffab7e4.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa77cb1e209131cea75602.md>)

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

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