# Table Question Answering on MMTU (BLEU/ROUGE-L)

Source: [Table Question Answering on MMTU (BLEU/ROUGE-L)](<https://www.sota2.com/research/sota/table-question-answering-on-mmtu-bleu-rouge-l>)

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

Data updated: 2026-02-26T05:37:00.398Z

Evaluation of model performance on the MMTU dataset for Table Question Answering, specifically measured using BLEU and ROUGE-L scores.

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 | Table Question Answering |
| Dataset (including any stated split) | MMTU |
| Metric | BLEU |
| Unit | PERCENT |
| Direction | Higher is better |
| Results loaded | 3 of 3 |
| 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

- [BLEU](<https://www.sota2.com/research/sota/table-question-answering-on-mmtu-bleu-rouge-l.md?metric=BLEU>): Higher is better; unit: PERCENT.

- [ROUGE-L](<https://www.sota2.com/research/sota/table-question-answering-on-mmtu-bleu-rouge-l.md?metric=ROUGE-L>): Higher is better; unit: PERCENT.

## Recorded results

| Method | BLEU | Result date | Report paper |
| --- | --- | --- | --- |
| GPT-5-nano (RE-TAB RG) | 21.19 | Not recorded | 69802f5ca8b01e76cc8e1465 |
| GPT-4.1-nano (RE-TAB RG) | 20.65 | Not recorded | 69802f5ca8b01e76cc8e1465 |
| QWEN3-8B (RE-TAB AS) | 6.5 | Not recorded | 69802f5ca8b01e76cc8e1465 |

### GPT-5-nano (RE-TAB RG)

BLEU: 21.19. Result ID: 699616665698665ec4560659.

[Reporting paper](<https://www.sota2.com/research/paper/69802f5ca8b01e76cc8e1465.md>)

[Method origin paper](<https://www.sota2.com/research/paper/69802f5ca8b01e76cc8e1465.md>)

Recorded evaluation configuration: {"Workflow":"Tree-of-Table (Tools)","ST Reward":"Enabled","SR Reward":"RG"}. This record does not assert independent reproduction.

### GPT-4.1-nano (RE-TAB RG)

BLEU: 20.65. Result ID: 699616665698665ec456065a.

[Reporting paper](<https://www.sota2.com/research/paper/69802f5ca8b01e76cc8e1465.md>)

[Method origin paper](<https://www.sota2.com/research/paper/69802f5ca8b01e76cc8e1465.md>)

Recorded evaluation configuration: {"Workflow":"Tree-of-Table (Tools)","ST Reward":"Enabled","SR Reward":"RG"}. This record does not assert independent reproduction.

### QWEN3-8B (RE-TAB AS)

BLEU: 6.5. Result ID: 699616665698665ec456065b.

[Reporting paper](<https://www.sota2.com/research/paper/69802f5ca8b01e76cc8e1465.md>)

[Method origin paper](<https://www.sota2.com/research/paper/69802f5ca8b01e76cc8e1465.md>)

Recorded evaluation configuration: {"Workflow":"Chain-of-Table (Tools)","ST Reward":"Enabled","SR Reward":"AS"}. This record does not assert independent reproduction.

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