# 3D City Generation on 100 textual descriptions of cities

Source: [3D City Generation on 100 textual descriptions of cities](<https://www.sota2.com/research/sota/3d-city-generation-on-100-textual-descriptions-of-cities>)

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

Data updated: 2026-03-10T07:33:21.044Z

Evaluation of 3D city generation models based on 100 textual descriptions, assessed using human and GPT-5 win rates across specific fidelity and realism dimensions (e.g., Geometric Fidelity, Texture Clarity, Layout Coherence).

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 | 3D City Generation |
| Dataset (including any stated split) | 100 textual descriptions of cities |
| Metric | VQA Score |
| Unit | SCALAR |
| Direction | Higher is better |
| Results loaded | 8 of 8 |
| 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

- [VQA Score](<https://www.sota2.com/research/sota/3d-city-generation-on-100-textual-descriptions-of-cities.md?metric=VQA%20Score>): Higher is better; unit: SCALAR.

- [Geometric Fidelity (GPT-5 Win Rate)](<https://www.sota2.com/research/sota/3d-city-generation-on-100-textual-descriptions-of-cities.md?metric=Geometric%20Fidelity%20(GPT-5%20Win%20Rate)>): Higher is better; unit: PERCENT.

- [Geometric Fidelity (Human Win Rate)](<https://www.sota2.com/research/sota/3d-city-generation-on-100-textual-descriptions-of-cities.md?metric=Geometric%20Fidelity%20(Human%20Win%20Rate)>): Higher is better; unit: PERCENT.

- [Texture Clarity (GPT-5 Win Rate)](<https://www.sota2.com/research/sota/3d-city-generation-on-100-textual-descriptions-of-cities.md?metric=Texture%20Clarity%20(GPT-5%20Win%20Rate)>): Higher is better; unit: PERCENT.

- [Texture Clarity (Human Win Rate)](<https://www.sota2.com/research/sota/3d-city-generation-on-100-textual-descriptions-of-cities.md?metric=Texture%20Clarity%20(Human%20Win%20Rate)>): Higher is better; unit: PERCENT.

- [Layout Coherence (GPT-5 Win Rate)](<https://www.sota2.com/research/sota/3d-city-generation-on-100-textual-descriptions-of-cities.md?metric=Layout%20Coherence%20(GPT-5%20Win%20Rate)>): Higher is better; unit: PERCENT.

- [Layout Coherence (Human Win Rate)](<https://www.sota2.com/research/sota/3d-city-generation-on-100-textual-descriptions-of-cities.md?metric=Layout%20Coherence%20(Human%20Win%20Rate)>): Higher is better; unit: PERCENT.

- [Scene Coverage (GPT-5 Win Rate)](<https://www.sota2.com/research/sota/3d-city-generation-on-100-textual-descriptions-of-cities.md?metric=Scene%20Coverage%20(GPT-5%20Win%20Rate)>): Higher is better; unit: PERCENT.

- [Scene Coverage (Human Win Rate)](<https://www.sota2.com/research/sota/3d-city-generation-on-100-textual-descriptions-of-cities.md?metric=Scene%20Coverage%20(Human%20Win%20Rate)>): Higher is better; unit: PERCENT.

- [Overall Realism (GPT-5 Win Rate)](<https://www.sota2.com/research/sota/3d-city-generation-on-100-textual-descriptions-of-cities.md?metric=Overall%20Realism%20(GPT-5%20Win%20Rate)>): Higher is better; unit: PERCENT.

- [Overall Realism (Human Win Rate)](<https://www.sota2.com/research/sota/3d-city-generation-on-100-textual-descriptions-of-cities.md?metric=Overall%20Realism%20(Human%20Win%20Rate)>): Higher is better; unit: PERCENT.

## Recorded results

| Method | VQA Score | Result date | Report paper |
| --- | --- | --- | --- |
| Yo’City | 0.7151 | Not recorded | 692538a1d6b0dd334292f98f |
| Yo’City | 0.7151 | Not recorded | 692538a1d6b0dd334292f98f |
| Yo’City | 0.7151 | Not recorded | 692538a1d6b0dd334292f98f |
| Yo’City | 0.7151 | Not recorded | 692538a1d6b0dd334292f98f |
| SynCity | 0.6975 | Not recorded | 692538a1d6b0dd334292f98f |
| Hunyuan3D (API) | 0.6198 | Not recorded | 692538a1d6b0dd334292f98f |
| Trellis | 0.6189 | Not recorded | 692538a1d6b0dd334292f98f |
| CityCraft | 0.5639 | Not recorded | 692538a1d6b0dd334292f98f |

### Yo’City

VQA Score: 0.7151. Result ID: 69afc9286dbb7075b957c8c3.

[Reporting paper](<https://www.sota2.com/research/paper/692538a1d6b0dd334292f98f.md>)

[Method origin paper](<https://www.sota2.com/research/paper/692538a1d6b0dd334292f98f.md>)

Recorded evaluation configuration: {"Comparison Target":"Trellis"}. This record does not assert independent reproduction.

### Yo’City

VQA Score: 0.7151. Result ID: 69afc9286dbb7075b957c8c5.

[Reporting paper](<https://www.sota2.com/research/paper/692538a1d6b0dd334292f98f.md>)

[Method origin paper](<https://www.sota2.com/research/paper/692538a1d6b0dd334292f98f.md>)

Recorded evaluation configuration: {"Comparison Target":"Hunyuan3D (API)"}. This record does not assert independent reproduction.

### Yo’City

VQA Score: 0.7151. Result ID: 69afc9286dbb7075b957c8c7.

[Reporting paper](<https://www.sota2.com/research/paper/692538a1d6b0dd334292f98f.md>)

[Method origin paper](<https://www.sota2.com/research/paper/692538a1d6b0dd334292f98f.md>)

Recorded evaluation configuration: {"Comparison Target":"CityCraft"}. This record does not assert independent reproduction.

### Yo’City

VQA Score: 0.7151. Result ID: 69afc9286dbb7075b957c8c9.

[Reporting paper](<https://www.sota2.com/research/paper/692538a1d6b0dd334292f98f.md>)

[Method origin paper](<https://www.sota2.com/research/paper/692538a1d6b0dd334292f98f.md>)

Recorded evaluation configuration: {"Comparison Target":"SynCity"}. This record does not assert independent reproduction.

### SynCity

VQA Score: 0.6975. Result ID: 69afc9286dbb7075b957c8c8.

[Reporting paper](<https://www.sota2.com/research/paper/692538a1d6b0dd334292f98f.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa77811e209131cea73a5d.md>)

Recorded evaluation configuration: {"Comparison Target":"Yo’City"}. This record does not assert independent reproduction.

### Hunyuan3D (API)

VQA Score: 0.6198. Result ID: 69afc9286dbb7075b957c8c4.

[Reporting paper](<https://www.sota2.com/research/paper/692538a1d6b0dd334292f98f.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa79911e209131cea7e495.md>)

Recorded evaluation configuration: {"Comparison Target":"Yo’City"}. This record does not assert independent reproduction.

### Trellis

VQA Score: 0.6189. Result ID: 69afc9286dbb7075b957c8c2.

[Reporting paper](<https://www.sota2.com/research/paper/692538a1d6b0dd334292f98f.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa78f21e209131cea7b608.md>)

[Implementation](<https://trellis3d.github.io/>)

Recorded evaluation configuration: {"Comparison Target":"Yo’City"}. This record does not assert independent reproduction.

### CityCraft

VQA Score: 0.5639. Result ID: 69afc9286dbb7075b957c8c6.

[Reporting paper](<https://www.sota2.com/research/paper/692538a1d6b0dd334292f98f.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa73831e209131cea5cdae.md>)

Recorded evaluation configuration: {"Comparison Target":"Yo’City"}. This record does not assert independent reproduction.

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