# Tabular Data Generation on Beijing (test)

Source: [Tabular Data Generation on Beijing (test)](<https://www.sota2.com/research/sota/tabular-data-generation-on-beijing-test>)

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

Data updated: 2026-04-23T07:47:41.282Z

Evaluation of model performance on generating synthetic tabular data using the test split of the Beijing dataset, measured by metrics like MLE and rho.

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 | Tabular Data Generation |
| Dataset (including any stated split) | Beijing (test) |
| Metric | MLE |
| Unit | SCALAR |
| Direction | Lower is better |
| Results loaded | 29 of 29 |
| 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

- [MLE](<https://www.sota2.com/research/sota/tabular-data-generation-on-beijing-test.md?metric=MLE>): Lower is better; unit: SCALAR.

- [Spearman Rho](<https://www.sota2.com/research/sota/tabular-data-generation-on-beijing-test.md?metric=Spearman%20Rho>): Higher is better; unit: SCALAR.

- [RMSE](<https://www.sota2.com/research/sota/tabular-data-generation-on-beijing-test.md?metric=RMSE>): Lower is better; unit: SCALAR.

- [Alpha Precision](<https://www.sota2.com/research/sota/tabular-data-generation-on-beijing-test.md?metric=Alpha%20Precision>): Higher is better; unit: PERCENT.

## Recorded results

| Method | MLE | Result date | Report paper |
| --- | --- | --- | --- |
| Real | 0.423 | Not recorded | 68fa795e1e209131cea7d4f3 |
| TabSyn | 0.582 | Not recorded | 68fa795e1e209131cea7d4f3 |
| SMOTE | 0.593 | Not recorded | 68fa795e1e209131cea7d4f3 |
| GReaT | 0.653 | Not recorded | 68fa795e1e209131cea7d4f3 |
| DiffLM | 0.696 | Not recorded | 68fa795e1e209131cea7d4f3 |
| TVAE | 0.77 | Not recorded | 68fa795e1e209131cea7d4f3 |
| CoDi | 0.818 | Not recorded | 68fa795e1e209131cea7d4f3 |
| Mistral-7B ICL | 0.865 | Not recorded | 68fa795e1e209131cea7d4f3 |
| CTGAN | 0.902 | Not recorded | 68fa795e1e209131cea7d4f3 |
| GPT-4 ICL | 0.992 | Not recorded | 68fa795e1e209131cea7d4f3 |
| GOGGLE | 1.09 | Not recorded | 68fa795e1e209131cea7d4f3 |
| Qwen2.5-14B ICL | 1.289 | Not recorded | 68fa795e1e209131cea7d4f3 |

### Real

MLE: 0.423. Result ID: 698b4a058d67a41d0a6cd5f4.

[Reporting paper](<https://www.sota2.com/research/paper/68fa795e1e209131cea7d4f3.md>)

Recorded evaluation configuration: {"Source":"Original Data"}. This record does not assert independent reproduction.

### TabSyn

MLE: 0.582. Result ID: 698b4a058d67a41d0a6cd5fa.

[Reporting paper](<https://www.sota2.com/research/paper/68fa795e1e209131cea7d4f3.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa73261e209131cea5ab3b.md>)

Recorded evaluation configuration: {"Architecture":"Diffusion-based"}. This record does not assert independent reproduction.

### SMOTE

MLE: 0.593. Result ID: 698b4a058d67a41d0a6cd5f5.

[Reporting paper](<https://www.sota2.com/research/paper/68fa795e1e209131cea7d4f3.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa60691e209131ce9ef1af.md>)

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

### GReaT

MLE: 0.653. Result ID: 698b4a058d67a41d0a6cd5fe.

[Reporting paper](<https://www.sota2.com/research/paper/68fa795e1e209131cea7d4f3.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6f291e209131cea442a2.md>)

Recorded evaluation configuration: {"Architecture":"LLM-based"}. This record does not assert independent reproduction.

### DiffLM

MLE: 0.696. Result ID: 698b4a058d67a41d0a6cd5ff.

[Reporting paper](<https://www.sota2.com/research/paper/68fa795e1e209131cea7d4f3.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa795e1e209131cea7d4f3.md>)

[Implementation](<https://github.com/bytedance/DiffLM>)

Recorded evaluation configuration: {"Architecture":"Diffusion-based","Decoder":"Mistral-v0.3"}. This record does not assert independent reproduction.

### TVAE

MLE: 0.77. Result ID: 698b4a058d67a41d0a6cd5f7.

[Reporting paper](<https://www.sota2.com/research/paper/68fa795e1e209131cea7d4f3.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa66901e209131cea10e19.md>)

[Implementation](<https://github.com/DAI-Lab/CTGAN>)

Recorded evaluation configuration: {"Architecture":"VAE-based"}. This record does not assert independent reproduction.

### CoDi

MLE: 0.818. Result ID: 698b4a058d67a41d0a6cd5f9.

[Reporting paper](<https://www.sota2.com/research/paper/68fa795e1e209131cea7d4f3.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa70b21e209131cea4ca44.md>)

Recorded evaluation configuration: {"Architecture":"Diffusion-based"}. This record does not assert independent reproduction.

### Mistral-7B ICL

MLE: 0.865. Result ID: 698b4a058d67a41d0a6cd5fc.

[Reporting paper](<https://www.sota2.com/research/paper/68fa795e1e209131cea7d4f3.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa70d61e209131cea4d909.md>)

[Implementation](<https://github.com/mistralai/mistral-src>)

Recorded evaluation configuration: {"Protocol":"5-shot ICL"}. This record does not assert independent reproduction.

### CTGAN

MLE: 0.902. Result ID: 698b4a058d67a41d0a6cd5f6.

[Reporting paper](<https://www.sota2.com/research/paper/68fa795e1e209131cea7d4f3.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa66901e209131cea10e19.md>)

[Implementation](<https://github.com/DAI-Lab/CTGAN>)

Recorded evaluation configuration: {"Architecture":"GAN-based"}. This record does not assert independent reproduction.

### GPT-4 ICL

MLE: 0.992. Result ID: 698b4a058d67a41d0a6cd5fb.

[Reporting paper](<https://www.sota2.com/research/paper/68fa795e1e209131cea7d4f3.md>)

Recorded evaluation configuration: {"Protocol":"5-shot ICL"}. This record does not assert independent reproduction.

### GOGGLE

MLE: 1.09. Result ID: 698b4a058d67a41d0a6cd5f8.

[Reporting paper](<https://www.sota2.com/research/paper/68fa795e1e209131cea7d4f3.md>)

Recorded evaluation configuration: {"Architecture":"VAE-based"}. This record does not assert independent reproduction.

### Qwen2.5-14B ICL

MLE: 1.289. Result ID: 698b4a058d67a41d0a6cd5fd.

[Reporting paper](<https://www.sota2.com/research/paper/68fa795e1e209131cea7d4f3.md>)

Recorded evaluation configuration: {"Protocol":"5-shot ICL"}. This record does not assert independent reproduction.

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