# Text-to-Image on GenEval 11 (test)

Source: [Text-to-Image on GenEval 11 (test)](<https://www.sota2.com/research/sota/text-to-image-on-geneval-11-test>)

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

Data updated: 2026-02-26T05:34:26.946Z

Evaluation of Text-to-Image generation models on the GenEval 11 test set, focusing on compositional accuracy metrics including object presence, counting, color fidelity, and spatial positioning.

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 | Text-to-Image |
| Dataset (including any stated split) | GenEval 11 (test) |
| Metric | Accuracy (Single Obj) |
| Unit | PERCENT |
| Direction | Higher 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

- [Accuracy (Single Obj)](<https://www.sota2.com/research/sota/text-to-image-on-geneval-11-test.md?metric=Accuracy%20(Single%20Obj)>): Higher is better; unit: PERCENT.

- [Accuracy (Two Obj.)](<https://www.sota2.com/research/sota/text-to-image-on-geneval-11-test.md?metric=Accuracy%20(Two%20Obj.)>): Higher is better; unit: PERCENT.

- [Counting Accuracy](<https://www.sota2.com/research/sota/text-to-image-on-geneval-11-test.md?metric=Counting%20Accuracy>): Higher is better; unit: PERCENT.

- [Color Fidelity](<https://www.sota2.com/research/sota/text-to-image-on-geneval-11-test.md?metric=Color%20Fidelity>): Higher is better; unit: PERCENT.

- [Spatial Positioning Accuracy](<https://www.sota2.com/research/sota/text-to-image-on-geneval-11-test.md?metric=Spatial%20Positioning%20Accuracy>): Higher is better; unit: PERCENT.

- [Color Attribute Accuracy](<https://www.sota2.com/research/sota/text-to-image-on-geneval-11-test.md?metric=Color%20Attribute%20Accuracy>): Higher is better; unit: PERCENT.

- [Overall Score](<https://www.sota2.com/research/sota/text-to-image-on-geneval-11-test.md?metric=Overall%20Score>): Higher is better; unit: PERCENT.

## Recorded results

| Method | Accuracy (Single Obj) | Result date | Report paper |
| --- | --- | --- | --- |
| SD 3.5 M | 100 | Not recorded | 692689c5d6b0dd3342930642 |
| Individual Expert (Flow-GRPO) | 100 | Not recorded | 692689c5d6b0dd3342930642 |
| Individual Expert (Flow-GRPO) | 100 | Not recorded | 692689c5d6b0dd3342930642 |
| Individual Expert (Flow-GRPO) | 100 | Not recorded | 692689c5d6b0dd3342930642 |
| Rewarded Soup | 100 | Not recorded | 692689c5d6b0dd3342930642 |
| MORL-D | 100 | Not recorded | 692689c5d6b0dd3342930642 |
| MORL-DR | 100 | Not recorded | 692689c5d6b0dd3342930642 |
| MapReduce LoRA | 100 | Not recorded | 692689c5d6b0dd3342930642 |
| MapReduce LoRA + RaTE | 100 | Not recorded | 692689c5d6b0dd3342930642 |
| Individual Expert (Flow-GRPO) | 100 | Not recorded | 692689c5d6b0dd3342930642 |
| Individual Expert (Flow-GRPO) | 100 | Not recorded | 692689c5d6b0dd3342930642 |
| Individual Expert (Flow-GRPO) | 100 | Not recorded | 692689c5d6b0dd3342930642 |
| MORL-D | 100 | Not recorded | 692689c5d6b0dd3342930642 |
| MORL-DR | 100 | Not recorded | 692689c5d6b0dd3342930642 |
| MapReduce LoRA | 100 | Not recorded | 692689c5d6b0dd3342930642 |
| SD 3 M | 99 | Not recorded | 692689c5d6b0dd3342930642 |
| SD 3 M + CaPO | 99 | Not recorded | 692689c5d6b0dd3342930642 |
| FLUX.1-dev | 99 | Not recorded | 692689c5d6b0dd3342930642 |
| Rewarded Soup | 99 | Not recorded | 692689c5d6b0dd3342930642 |

### SD 3.5 M

Accuracy (Single Obj): 100. Result ID: 699a3c8b5698665ec4c30e28.

[Reporting paper](<https://www.sota2.com/research/paper/692689c5d6b0dd3342930642.md>)

Recorded evaluation configuration: {"Backbone":"SD 3.5 M","Reward":"None"}. This record does not assert independent reproduction.

### Individual Expert (Flow-GRPO)

Accuracy (Single Obj): 100. Result ID: 699a3c8b5698665ec4c30e29.

[Reporting paper](<https://www.sota2.com/research/paper/692689c5d6b0dd3342930642.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa7a061e209131cea8093f.md>)

[Implementation](<https://github.com/yifan123/flow_grpo>)

Recorded evaluation configuration: {"Backbone":"SD 3.5 M","Reward":"GenEval \[11\]"}. This record does not assert independent reproduction.

### Individual Expert (Flow-GRPO)

Accuracy (Single Obj): 100. Result ID: 699a3c8b5698665ec4c30e2a.

[Reporting paper](<https://www.sota2.com/research/paper/692689c5d6b0dd3342930642.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa7a061e209131cea8093f.md>)

[Implementation](<https://github.com/yifan123/flow_grpo>)

Recorded evaluation configuration: {"Backbone":"SD 3.5 M","Reward":"PickScore \[15\]"}. This record does not assert independent reproduction.

### Individual Expert (Flow-GRPO)

Accuracy (Single Obj): 100. Result ID: 699a3c8b5698665ec4c30e2b.

[Reporting paper](<https://www.sota2.com/research/paper/692689c5d6b0dd3342930642.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa7a061e209131cea8093f.md>)

[Implementation](<https://github.com/yifan123/flow_grpo>)

Recorded evaluation configuration: {"Backbone":"SD 3.5 M","Reward":"OCR \[6\]"}. This record does not assert independent reproduction.

### Rewarded Soup

Accuracy (Single Obj): 100. Result ID: 699a3c8b5698665ec4c30e2c.

[Reporting paper](<https://www.sota2.com/research/paper/692689c5d6b0dd3342930642.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa70e61e209131cea4dcbb.md>)

Recorded evaluation configuration: {"Backbone":"SD 3.5 M","Reward":"\[11\] + \[15\] + \[6\]"}. This record does not assert independent reproduction.

### MORL-D

Accuracy (Single Obj): 100. Result ID: 699a3c8b5698665ec4c30e2d.

[Reporting paper](<https://www.sota2.com/research/paper/692689c5d6b0dd3342930642.md>)

Recorded evaluation configuration: {"Backbone":"SD 3.5 M","Reward":"\[11\] + \[15\] + \[6\]"}. This record does not assert independent reproduction.

### MORL-DR

Accuracy (Single Obj): 100. Result ID: 699a3c8b5698665ec4c30e2e.

[Reporting paper](<https://www.sota2.com/research/paper/692689c5d6b0dd3342930642.md>)

Recorded evaluation configuration: {"Backbone":"SD 3.5 M","Reward":"\[11\] + \[15\] + \[6\]"}. This record does not assert independent reproduction.

### MapReduce LoRA

Accuracy (Single Obj): 100. Result ID: 699a3c8b5698665ec4c30e2f.

[Reporting paper](<https://www.sota2.com/research/paper/692689c5d6b0dd3342930642.md>)

[Method origin paper](<https://www.sota2.com/research/paper/692689c5d6b0dd3342930642.md>)

Recorded evaluation configuration: {"Backbone":"SD 3.5 M","Reward":"\[11\] + \[15\] + \[6\]"}. This record does not assert independent reproduction.

### MapReduce LoRA + RaTE

Accuracy (Single Obj): 100. Result ID: 699a3c8b5698665ec4c30e30.

[Reporting paper](<https://www.sota2.com/research/paper/692689c5d6b0dd3342930642.md>)

[Method origin paper](<https://www.sota2.com/research/paper/692689c5d6b0dd3342930642.md>)

Recorded evaluation configuration: {"Backbone":"SD 3.5 M","Reward":"\[11\] + \[15\] + \[6\]"}. This record does not assert independent reproduction.

### Individual Expert (Flow-GRPO)

Accuracy (Single Obj): 100. Result ID: 699a3c8b5698665ec4c30e32.

[Reporting paper](<https://www.sota2.com/research/paper/692689c5d6b0dd3342930642.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa7a061e209131cea8093f.md>)

[Implementation](<https://github.com/yifan123/flow_grpo>)

Recorded evaluation configuration: {"Backbone":"FLUX.1-dev","Reward":"GenEval \[11\]"}. This record does not assert independent reproduction.

### Individual Expert (Flow-GRPO)

Accuracy (Single Obj): 100. Result ID: 699a3c8b5698665ec4c30e33.

[Reporting paper](<https://www.sota2.com/research/paper/692689c5d6b0dd3342930642.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa7a061e209131cea8093f.md>)

[Implementation](<https://github.com/yifan123/flow_grpo>)

Recorded evaluation configuration: {"Backbone":"FLUX.1-dev","Reward":"PickScore \[15\]"}. This record does not assert independent reproduction.

### Individual Expert (Flow-GRPO)

Accuracy (Single Obj): 100. Result ID: 699a3c8b5698665ec4c30e34.

[Reporting paper](<https://www.sota2.com/research/paper/692689c5d6b0dd3342930642.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa7a061e209131cea8093f.md>)

[Implementation](<https://github.com/yifan123/flow_grpo>)

Recorded evaluation configuration: {"Backbone":"FLUX.1-dev","Reward":"OCR \[6\]"}. This record does not assert independent reproduction.

### MORL-D

Accuracy (Single Obj): 100. Result ID: 699a3c8b5698665ec4c30e36.

[Reporting paper](<https://www.sota2.com/research/paper/692689c5d6b0dd3342930642.md>)

Recorded evaluation configuration: {"Backbone":"FLUX.1-dev","Reward":"\[11\] + \[15\] + \[6\]"}. This record does not assert independent reproduction.

### MORL-DR

Accuracy (Single Obj): 100. Result ID: 699a3c8b5698665ec4c30e37.

[Reporting paper](<https://www.sota2.com/research/paper/692689c5d6b0dd3342930642.md>)

Recorded evaluation configuration: {"Backbone":"FLUX.1-dev","Reward":"\[11\] + \[15\] + \[6\]"}. This record does not assert independent reproduction.

### MapReduce LoRA

Accuracy (Single Obj): 100. Result ID: 699a3c8b5698665ec4c30e38.

[Reporting paper](<https://www.sota2.com/research/paper/692689c5d6b0dd3342930642.md>)

[Method origin paper](<https://www.sota2.com/research/paper/692689c5d6b0dd3342930642.md>)

Recorded evaluation configuration: {"Backbone":"FLUX.1-dev","Reward":"\[11\] + \[15\] + \[6\]"}. This record does not assert independent reproduction.

### SD 3 M

Accuracy (Single Obj): 99. Result ID: 699a3c8b5698665ec4c30e26.

[Reporting paper](<https://www.sota2.com/research/paper/692689c5d6b0dd3342930642.md>)

Recorded evaluation configuration: {"Backbone":"SD 3 M","Reward":"None"}. This record does not assert independent reproduction.

### SD 3 M + CaPO

Accuracy (Single Obj): 99. Result ID: 699a3c8b5698665ec4c30e27.

[Reporting paper](<https://www.sota2.com/research/paper/692689c5d6b0dd3342930642.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa7b9f1e209131cea89667.md>)

Recorded evaluation configuration: {"Backbone":"SD 3 M","Reward":"VQAScore \[21\] + MPS \[43\] + VILA \[14\]"}. This record does not assert independent reproduction.

### FLUX.1-dev

Accuracy (Single Obj): 99. Result ID: 699a3c8b5698665ec4c30e31.

[Reporting paper](<https://www.sota2.com/research/paper/692689c5d6b0dd3342930642.md>)

Recorded evaluation configuration: {"Backbone":"FLUX.1-dev","Reward":"None"}. This record does not assert independent reproduction.

### Rewarded Soup

Accuracy (Single Obj): 99. Result ID: 699a3c8b5698665ec4c30e35.

[Reporting paper](<https://www.sota2.com/research/paper/692689c5d6b0dd3342930642.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa70e61e209131cea4dcbb.md>)

Recorded evaluation configuration: {"Backbone":"FLUX.1-dev","Reward":"\[11\] + \[15\] + \[6\]"}. This record does not assert independent reproduction.

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