# MBR decoding on WMT En-Ja 2024 (test)

Source: [MBR decoding on WMT En-Ja 2024 (test)](<https://www.sota2.com/research/sota/mbr-decoding-on-wmt-en-ja-2024-test>)

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

Data updated: 2026-02-26T06:09:03.204Z

Evaluation of Minimum Bayes Risk (MBR) decoding performance on the WMT 2024 English-to-Japanese test set, measured using XCOMET-XL, CometKiwi-XL, and MetricX-XL.

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 | MBR decoding |
| Dataset (including any stated split) | WMT En-Ja 2024 (test) |
| Metric | XCOMET-XL |
| Unit | SCALAR |
| Direction | Higher is better |
| Results loaded | 4 of 4 |
| 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

- [XCOMET-XL](<https://www.sota2.com/research/sota/mbr-decoding-on-wmt-en-ja-2024-test.md?metric=XCOMET-XL>): Higher is better; unit: SCALAR.

- [CometKiwi-XL](<https://www.sota2.com/research/sota/mbr-decoding-on-wmt-en-ja-2024-test.md?metric=CometKiwi-XL>): Higher is better; unit: SCALAR.

- [MetricX-XL](<https://www.sota2.com/research/sota/mbr-decoding-on-wmt-en-ja-2024-test.md?metric=MetricX-XL>): Higher is better; unit: SCALAR.

## Recorded results

| Method | XCOMET-XL | Result date | Report paper |
| --- | --- | --- | --- |
| PEAR (full) | 0.81 | Not recorded | 69784706a2b28e761d87c91d |
| PEAR (sym.) | 0.809 | Not recorded | 69784706a2b28e761d87c91d |
| COMET-22 | 0.798 | Not recorded | 69784706a2b28e761d87c91d |
| BLEURT-20 | 0.796 | Not recorded | 69784706a2b28e761d87c91d |

### PEAR (full)

XCOMET-XL: 0.81. Result ID: 69966df65698665ec4b4d4c7.

[Reporting paper](<https://www.sota2.com/research/paper/69784706a2b28e761d87c91d.md>)

[Method origin paper](<https://www.sota2.com/research/paper/69784706a2b28e761d87c91d.md>)

Recorded evaluation configuration: {"Implementation":"Full utility matrix"}. This record does not assert independent reproduction.

### PEAR (sym.)

XCOMET-XL: 0.809. Result ID: 69966df65698665ec4b4d4c8.

[Reporting paper](<https://www.sota2.com/research/paper/69784706a2b28e761d87c91d.md>)

[Method origin paper](<https://www.sota2.com/research/paper/69784706a2b28e761d87c91d.md>)

Recorded evaluation configuration: {"Implementation":"Antisymmetric utility matrix"}. This record does not assert independent reproduction.

### COMET-22

XCOMET-XL: 0.798. Result ID: 69966df65698665ec4b4d4c9.

[Reporting paper](<https://www.sota2.com/research/paper/69784706a2b28e761d87c91d.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa688c1e209131cea1cdb6.md>)

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

### BLEURT-20

XCOMET-XL: 0.796. Result ID: 69966df65698665ec4b4d4ca.

[Reporting paper](<https://www.sota2.com/research/paper/69784706a2b28e761d87c91d.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa688e1e209131cea1cf4f.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.