# Text-to-Image Faithfulness on DSG

Source: [Text-to-Image Faithfulness on DSG](<https://www.sota2.com/research/sota/text-to-image-faithfulness-on-dsg>)

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

Data updated: 2026-07-01T08:24:36.857Z

Evaluation of Text-to-Image model faithfulness on the DSG dataset, specifically measured by the 'Faithfulness yes-ratio'.

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 Faithfulness |
| Dataset (including any stated split) | DSG |
| Metric | Faithfulness Ratio |
| Unit | PERCENT |
| Direction | Higher is better |
| Results loaded | 11 of 11 |
| 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

- [Faithfulness Ratio](<https://www.sota2.com/research/sota/text-to-image-faithfulness-on-dsg.md?metric=Faithfulness%20Ratio>): Higher is better; unit: PERCENT.

## Recorded results

| Method | Faithfulness Ratio | Result date | Report paper |
| --- | --- | --- | --- |
| gemini-3-pro-image-preview-2k | 94.6 | Not recorded | 6a449f15e0224bd4f49931c2 |
| gpt-image-1.5-high-fidelity | 94.2 | Not recorded | 6a449f15e0224bd4f49931c2 |
| gpt-image-1 | 93.8 | Not recorded | 6a449f15e0224bd4f49931c2 |
| grok-imagine-image-20260306 | 93.4 | Not recorded | 6a449f15e0224bd4f49931c2 |
| imagen-4.0-ultra-generate-001 | 92.8 | Not recorded | 6a449f15e0224bd4f49931c2 |
| gemini-2.5-flash-image (nano-banana) | 92.1 | Not recorded | 6a449f15e0224bd4f49931c2 |
| wan2.6-t2i-v2 | 91.7 | Not recorded | 6a449f15e0224bd4f49931c2 |
| recraft-v4 | 91.4 | Not recorded | 6a449f15e0224bd4f49931c2 |
| imagen-4.0-generate-001 | 91.3 | Not recorded | 6a449f15e0224bd4f49931c2 |
| ideogram-v3-quality | 85.5 | Not recorded | 6a449f15e0224bd4f49931c2 |
| hunyuan-image-3.0-fal | 83.7 | Not recorded | 6a449f15e0224bd4f49931c2 |

### gemini-3-pro-image-preview-2k

Faithfulness Ratio: 94.6. Result ID: 6a44cec473d097d845bf5be3.

[Reporting paper](<https://www.sota2.com/research/paper/6a449f15e0224bd4f49931c2.md>)

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

### gpt-image-1.5-high-fidelity

Faithfulness Ratio: 94.2. Result ID: 6a44cec473d097d845bf5be5.

[Reporting paper](<https://www.sota2.com/research/paper/6a449f15e0224bd4f49931c2.md>)

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

### gpt-image-1

Faithfulness Ratio: 93.8. Result ID: 6a44cec473d097d845bf5be9.

[Reporting paper](<https://www.sota2.com/research/paper/6a449f15e0224bd4f49931c2.md>)

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

### grok-imagine-image-20260306

Faithfulness Ratio: 93.4. Result ID: 6a44cec473d097d845bf5be4.

[Reporting paper](<https://www.sota2.com/research/paper/6a449f15e0224bd4f49931c2.md>)

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

### imagen-4.0-ultra-generate-001

Faithfulness Ratio: 92.8. Result ID: 6a44cec473d097d845bf5bea.

[Reporting paper](<https://www.sota2.com/research/paper/6a449f15e0224bd4f49931c2.md>)

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

### gemini-2.5-flash-image (nano-banana)

Faithfulness Ratio: 92.1. Result ID: 6a44cec473d097d845bf5be8.

[Reporting paper](<https://www.sota2.com/research/paper/6a449f15e0224bd4f49931c2.md>)

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

### wan2.6-t2i-v2

Faithfulness Ratio: 91.7. Result ID: 6a44cec473d097d845bf5be7.

[Reporting paper](<https://www.sota2.com/research/paper/6a449f15e0224bd4f49931c2.md>)

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

### recraft-v4

Faithfulness Ratio: 91.4. Result ID: 6a44cec473d097d845bf5be6.

[Reporting paper](<https://www.sota2.com/research/paper/6a449f15e0224bd4f49931c2.md>)

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

### imagen-4.0-generate-001

Faithfulness Ratio: 91.3. Result ID: 6a44cec473d097d845bf5beb.

[Reporting paper](<https://www.sota2.com/research/paper/6a449f15e0224bd4f49931c2.md>)

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

### ideogram-v3-quality

Faithfulness Ratio: 85.5. Result ID: 6a44cec473d097d845bf5bed.

[Reporting paper](<https://www.sota2.com/research/paper/6a449f15e0224bd4f49931c2.md>)

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

### hunyuan-image-3.0-fal

Faithfulness Ratio: 83.7. Result ID: 6a44cec473d097d845bf5bec.

[Reporting paper](<https://www.sota2.com/research/paper/6a449f15e0224bd4f49931c2.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa7bb51e209131cea8a0c1.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.