# Text-to-Image Generation on DrawBench (PickScore, ClipScore, HPSv2.1, Aesthetic, ImgRwd, UniRwd)

Source: [Text-to-Image Generation on DrawBench (PickScore, ClipScore, HPSv2.1, Aesthetic, ImgRwd, UniRwd)](<https://www.sota2.com/research/sota/text-to-image-generation-on-drawbench-pickscore-clipscore-hpsv2-1-aesthetic-imgrwd-unirwd>)

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

Data updated: 2026-04-28T07:56:42.187Z

Comprehensive evaluation of Text-to-Image generation models on DrawBench, utilizing a suite of metrics covering perceptual quality (PickScore, ClipScore), fidelity (HPSv2.1), aesthetic appeal (Aesthetic), and learned reward signals (ImgRwd, UniRwd).

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 Generation |
| Dataset (including any stated split) | DrawBench |
| Metric | PickScore |
| Unit | SCALAR |
| Direction | Higher is better |
| Results loaded | 32 of 32 |
| 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

- [PickScore](<https://www.sota2.com/research/sota/text-to-image-generation-on-drawbench-pickscore-clipscore-hpsv2-1-aesthetic-imgrwd-unirwd.md?metric=PickScore>): Higher is better; unit: SCALAR.

- [ClipScore](<https://www.sota2.com/research/sota/text-to-image-generation-on-drawbench-pickscore-clipscore-hpsv2-1-aesthetic-imgrwd-unirwd.md?metric=ClipScore>): Higher is better; unit: SCALAR.

- [HPSv2.1](<https://www.sota2.com/research/sota/text-to-image-generation-on-drawbench-pickscore-clipscore-hpsv2-1-aesthetic-imgrwd-unirwd.md?metric=HPSv2.1>): Higher is better; unit: SCALAR.

- [Aesthetic Score](<https://www.sota2.com/research/sota/text-to-image-generation-on-drawbench-pickscore-clipscore-hpsv2-1-aesthetic-imgrwd-unirwd.md?metric=Aesthetic%20Score>): Higher is better; unit: SCALAR.

- [ImgRwd](<https://www.sota2.com/research/sota/text-to-image-generation-on-drawbench-pickscore-clipscore-hpsv2-1-aesthetic-imgrwd-unirwd.md?metric=ImgRwd>): Higher is better; unit: SCALAR.

- [UniRwd](<https://www.sota2.com/research/sota/text-to-image-generation-on-drawbench-pickscore-clipscore-hpsv2-1-aesthetic-imgrwd-unirwd.md?metric=UniRwd>): Higher is better; unit: SCALAR.

## Recorded results

| Method | PickScore | Result date | Report paper |
| --- | --- | --- | --- |
| DiffusionNFT | 23.8 | Not recorded | 68fa7b731e209131cea88901 |
| FlowGRPO | 23.5 | Not recorded | 68fa7b731e209131cea88901 |
| SD3.5-L | 22.91 | Not recorded | 68fa7b731e209131cea88901 |
| FLUX.1-Dev | 22.84 | Not recorded | 68fa7b731e209131cea88901 |
| Ctrl-Z‡ | 22.62 | Not recorded | 68fa7ba61e209131cea89a65 |
| SOP-4 | 22.58 | Not recorded | 68fa7ba61e209131cea89a65 |
| FlowGRPO | 22.51 | Not recorded | 68fa7b731e209131cea88901 |
| Ctrl-Z† | 22.51 | Not recorded | 68fa7ba61e209131cea89a65 |
| SOP-1 | 22.5 | Not recorded | 68fa7ba61e209131cea89a65 |
| Resampling | 22.49 | Not recorded | 68fa7ba61e209131cea89a65 |
| Z-Sampling | 22.49 | Not recorded | 68fa7ba61e209131cea89a65 |
| DDIM | 22.47 | Not recorded | 68fa7ba61e209131cea89a65 |
| SD-XL | 22.42 | Not recorded | 68fa7b731e209131cea88901 |
| FlowGRPO | 22.41 | Not recorded | 68fa7b731e209131cea88901 |
| SD3.5-M + CFG | 22.34 | Not recorded | 68fa7b731e209131cea88901 |
| SOP-4 | 21.89 | Not recorded | 68fa7ba61e209131cea89a65 |
| Ctrl-Z‡ | 21.84 | Not recorded | 68fa7ba61e209131cea89a65 |
| Ctrl-Z† | 21.68 | Not recorded | 68fa7ba61e209131cea89a65 |
| SOP-1 | 21.64 | Not recorded | 68fa7ba61e209131cea89a65 |
| Resampling | 21.52 | Not recorded | 68fa7ba61e209131cea89a65 |
| Z-Sampling | 21.47 | Not recorded | 68fa7ba61e209131cea89a65 |
| DDIM | 21.39 | Not recorded | 68fa7ba61e209131cea89a65 |
| SD3.5-M | 20.51 | Not recorded | 68fa7b731e209131cea88901 |
| SD 3.5 M + DiffusionNFT | 0.238 | Not recorded | 69f03ef155a3c32af89f9f42 |
| SD 3.5 M + FlowGRPO | 0.235 | Not recorded | 69f03ef155a3c32af89f9f42 |
| SD 3.5 M + V-GRPO | 0.235 | Not recorded | 69f03ef155a3c32af89f9f42 |
| SD 3.5 L | 0.2291 | Not recorded | 69f03ef155a3c32af89f9f42 |
| SD 3.5 M + FlowGRPO | 0.2251 | Not recorded | 69f03ef155a3c32af89f9f42 |
| SD XL | 0.2242 | Not recorded | 69f03ef155a3c32af89f9f42 |
| SD 3.5 M + FlowGRPO | 0.2241 | Not recorded | 69f03ef155a3c32af89f9f42 |
| SD 3.5 M + CFG | 0.2234 | Not recorded | 69f03ef155a3c32af89f9f42 |
| SD 3.5 M (w/o CFG) | 0.2051 | Not recorded | 69f03ef155a3c32af89f9f42 |

### DiffusionNFT

PickScore: 23.8. Result ID: 699a9f3a5698665ec42b1dee.

[Reporting paper](<https://www.sota2.com/research/paper/68fa7b731e209131cea88901.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa7b731e209131cea88901.md>)

Recorded evaluation configuration: {"#Iter":"1.7k","Resolution":"512x512"}. This record does not assert independent reproduction.

### FlowGRPO

PickScore: 23.5. Result ID: 699a9f3a5698665ec42b1ded.

[Reporting paper](<https://www.sota2.com/research/paper/68fa7b731e209131cea88901.md>)

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

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

Recorded evaluation configuration: {"#Iter":"4k","Resolution":"512x512"}. This record does not assert independent reproduction.

### SD3.5-L

PickScore: 22.91. Result ID: 699a9f3a5698665ec42b1de7.

[Reporting paper](<https://www.sota2.com/research/paper/68fa7b731e209131cea88901.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa72621e209131cea56163.md>)

Recorded evaluation configuration: {"Resolution":"1024x1024"}. This record does not assert independent reproduction.

### FLUX.1-Dev

PickScore: 22.84. Result ID: 699a9f3a5698665ec42b1de8.

[Reporting paper](<https://www.sota2.com/research/paper/68fa7b731e209131cea88901.md>)

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

### Ctrl-Z‡

PickScore: 22.62. Result ID: 69c2961847f34ce924aafc1a.

[Reporting paper](<https://www.sota2.com/research/paper/68fa7ba61e209131cea89a65.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa7ba61e209131cea89a65.md>)

Recorded evaluation configuration: {"Backbone":"Hunyuan-DiT","NFEs":"8.79"}. This record does not assert independent reproduction.

### SOP-4

PickScore: 22.58. Result ID: 69c2961847f34ce924aafc18.

[Reporting paper](<https://www.sota2.com/research/paper/68fa7ba61e209131cea89a65.md>)

Recorded evaluation configuration: {"Backbone":"Hunyuan-DiT","NFEs":"9.00"}. This record does not assert independent reproduction.

### FlowGRPO

PickScore: 22.51. Result ID: 699a9f3a5698665ec42b1deb.

[Reporting paper](<https://www.sota2.com/research/paper/68fa7b731e209131cea88901.md>)

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

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

Recorded evaluation configuration: {"#Iter":"\>5k","Resolution":"512x512"}. This record does not assert independent reproduction.

### Ctrl-Z†

PickScore: 22.51. Result ID: 69c2961847f34ce924aafc19.

[Reporting paper](<https://www.sota2.com/research/paper/68fa7ba61e209131cea89a65.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa7ba61e209131cea89a65.md>)

Recorded evaluation configuration: {"Backbone":"Hunyuan-DiT","NFEs":"2.85"}. This record does not assert independent reproduction.

### SOP-1

PickScore: 22.5. Result ID: 69c2961847f34ce924aafc17.

[Reporting paper](<https://www.sota2.com/research/paper/68fa7ba61e209131cea89a65.md>)

Recorded evaluation configuration: {"Backbone":"Hunyuan-DiT","NFEs":"3.00"}. This record does not assert independent reproduction.

### Resampling

PickScore: 22.49. Result ID: 69c2961847f34ce924aafc15.

[Reporting paper](<https://www.sota2.com/research/paper/68fa7ba61e209131cea89a65.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6d5a1e209131cea39634.md>)

[Implementation](<https://www.sota2.com/git.io/RePaint>)

Recorded evaluation configuration: {"Backbone":"Hunyuan-DiT","NFEs":"2.00"}. This record does not assert independent reproduction.

### Z-Sampling

PickScore: 22.49. Result ID: 69c2961847f34ce924aafc16.

[Reporting paper](<https://www.sota2.com/research/paper/68fa7ba61e209131cea89a65.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa76331e209131cea6bce5.md>)

Recorded evaluation configuration: {"Backbone":"Hunyuan-DiT","NFEs":"3.00"}. This record does not assert independent reproduction.

### DDIM

PickScore: 22.47. Result ID: 69c2961847f34ce924aafc14.

[Reporting paper](<https://www.sota2.com/research/paper/68fa7ba61e209131cea89a65.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6d8b1e209131cea3a7fe.md>)

Recorded evaluation configuration: {"Backbone":"Hunyuan-DiT","NFEs":"1.00"}. This record does not assert independent reproduction.

### SD-XL

PickScore: 22.42. Result ID: 699a9f3a5698665ec42b1de6.

[Reporting paper](<https://www.sota2.com/research/paper/68fa7b731e209131cea88901.md>)

Recorded evaluation configuration: {"Resolution":"1024x1024"}. This record does not assert independent reproduction.

### FlowGRPO

PickScore: 22.41. Result ID: 699a9f3a5698665ec42b1dec.

[Reporting paper](<https://www.sota2.com/research/paper/68fa7b731e209131cea88901.md>)

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

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

Recorded evaluation configuration: {"#Iter":"2k","Resolution":"512x512"}. This record does not assert independent reproduction.

### SD3.5-M + CFG

PickScore: 22.34. Result ID: 699a9f3a5698665ec42b1dea.

[Reporting paper](<https://www.sota2.com/research/paper/68fa7b731e209131cea88901.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa72621e209131cea56163.md>)

Recorded evaluation configuration: {"CFG":"true","Resolution":"512x512"}. This record does not assert independent reproduction.

### SOP-4

PickScore: 21.89. Result ID: 69c2961847f34ce924aafc11.

[Reporting paper](<https://www.sota2.com/research/paper/68fa7ba61e209131cea89a65.md>)

Recorded evaluation configuration: {"Backbone":"Stable Diffusion 2.1","NFEs":"9.00"}. This record does not assert independent reproduction.

### Ctrl-Z‡

PickScore: 21.84. Result ID: 69c2961847f34ce924aafc13.

[Reporting paper](<https://www.sota2.com/research/paper/68fa7ba61e209131cea89a65.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa7ba61e209131cea89a65.md>)

Recorded evaluation configuration: {"Backbone":"Stable Diffusion 2.1","NFEs":"7.72"}. This record does not assert independent reproduction.

### Ctrl-Z†

PickScore: 21.68. Result ID: 69c2961847f34ce924aafc12.

[Reporting paper](<https://www.sota2.com/research/paper/68fa7ba61e209131cea89a65.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa7ba61e209131cea89a65.md>)

Recorded evaluation configuration: {"Backbone":"Stable Diffusion 2.1","NFEs":"2.77"}. This record does not assert independent reproduction.

### SOP-1

PickScore: 21.64. Result ID: 69c2961847f34ce924aafc10.

[Reporting paper](<https://www.sota2.com/research/paper/68fa7ba61e209131cea89a65.md>)

Recorded evaluation configuration: {"Backbone":"Stable Diffusion 2.1","NFEs":"3.00"}. This record does not assert independent reproduction.

### Resampling

PickScore: 21.52. Result ID: 69c2961847f34ce924aafc0e.

[Reporting paper](<https://www.sota2.com/research/paper/68fa7ba61e209131cea89a65.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6d5a1e209131cea39634.md>)

[Implementation](<https://www.sota2.com/git.io/RePaint>)

Recorded evaluation configuration: {"Backbone":"Stable Diffusion 2.1","NFEs":"2.00"}. This record does not assert independent reproduction.

### Z-Sampling

PickScore: 21.47. Result ID: 69c2961847f34ce924aafc0f.

[Reporting paper](<https://www.sota2.com/research/paper/68fa7ba61e209131cea89a65.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa76331e209131cea6bce5.md>)

Recorded evaluation configuration: {"Backbone":"Stable Diffusion 2.1","NFEs":"3.00"}. This record does not assert independent reproduction.

### DDIM

PickScore: 21.39. Result ID: 69c2961847f34ce924aafc0d.

[Reporting paper](<https://www.sota2.com/research/paper/68fa7ba61e209131cea89a65.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6d8b1e209131cea3a7fe.md>)

Recorded evaluation configuration: {"Backbone":"Stable Diffusion 2.1","NFEs":"1.00"}. This record does not assert independent reproduction.

### SD3.5-M

PickScore: 20.51. Result ID: 699a9f3a5698665ec42b1de9.

[Reporting paper](<https://www.sota2.com/research/paper/68fa7b731e209131cea88901.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa72621e209131cea56163.md>)

Recorded evaluation configuration: {"CFG":"false","Resolution":"512x512"}. This record does not assert independent reproduction.

### SD 3.5 M + DiffusionNFT

PickScore: 0.238. Result ID: 69f0683147f34ce924d6177b.

[Reporting paper](<https://www.sota2.com/research/paper/69f03ef155a3c32af89f9f42.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa7b731e209131cea88901.md>)

Recorded evaluation configuration: {"#Steps":"1.7K","NFE\_pi\_theta\_old":"40 + 40","NFE\_pi\_theta":"40"}. This record does not assert independent reproduction.

### SD 3.5 M + FlowGRPO

PickScore: 0.235. Result ID: 69f0683147f34ce924d6177a.

[Reporting paper](<https://www.sota2.com/research/paper/69f03ef155a3c32af89f9f42.md>)

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

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

Recorded evaluation configuration: {"#Steps":"4K","NFE\_pi\_theta\_old":"40","NFE\_pi\_theta":"40"}. This record does not assert independent reproduction.

### SD 3.5 M + V-GRPO

PickScore: 0.235. Result ID: 69f0683147f34ce924d6177c.

[Reporting paper](<https://www.sota2.com/research/paper/69f03ef155a3c32af89f9f42.md>)

[Method origin paper](<https://www.sota2.com/research/paper/69f03ef155a3c32af89f9f42.md>)

[Implementation](<https://github.com/tang-bd/v-grpo>)

Recorded evaluation configuration: {"#Steps":"580","NFE\_pi\_theta\_old":"40 + 6.9","NFE\_pi\_theta":"6.9"}. This record does not assert independent reproduction.

### SD 3.5 L

PickScore: 0.2291. Result ID: 69f0683147f34ce924d61775.

[Reporting paper](<https://www.sota2.com/research/paper/69f03ef155a3c32af89f9f42.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa72621e209131cea56163.md>)

Recorded evaluation configuration: {"Resolution":"1024×1024"}. This record does not assert independent reproduction.

### SD 3.5 M + FlowGRPO

PickScore: 0.2251. Result ID: 69f0683147f34ce924d61778.

[Reporting paper](<https://www.sota2.com/research/paper/69f03ef155a3c32af89f9f42.md>)

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

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

Recorded evaluation configuration: {"#Steps":"\>5K","NFE\_pi\_theta\_old":"40","NFE\_pi\_theta":"40"}. This record does not assert independent reproduction.

### SD XL

PickScore: 0.2242. Result ID: 69f0683147f34ce924d61774.

[Reporting paper](<https://www.sota2.com/research/paper/69f03ef155a3c32af89f9f42.md>)

Recorded evaluation configuration: {"Resolution":"1024×1024"}. This record does not assert independent reproduction.

### SD 3.5 M + FlowGRPO

PickScore: 0.2241. Result ID: 69f0683147f34ce924d61779.

[Reporting paper](<https://www.sota2.com/research/paper/69f03ef155a3c32af89f9f42.md>)

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

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

Recorded evaluation configuration: {"#Steps":"2K","NFE\_pi\_theta\_old":"40","NFE\_pi\_theta":"40"}. This record does not assert independent reproduction.

### SD 3.5 M + CFG

PickScore: 0.2234. Result ID: 69f0683147f34ce924d61777.

[Reporting paper](<https://www.sota2.com/research/paper/69f03ef155a3c32af89f9f42.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa72621e209131cea56163.md>)

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

### SD 3.5 M (w/o CFG)

PickScore: 0.2051. Result ID: 69f0683147f34ce924d61776.

[Reporting paper](<https://www.sota2.com/research/paper/69f03ef155a3c32af89f9f42.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa72621e209131cea56163.md>)

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

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