# Machine Translation Evaluation on WMT 21

Source: [Machine Translation Evaluation on WMT 21](<https://www.sota2.com/research/sota/machine-translation-evaluation-on-wmt-21>)

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

Data updated: 2026-03-11T08:42:17.972Z

Evaluation of machine translation quality on the WMT 21 dataset, reporting scores for specific language pairs such as en-ha and xh-zu.

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 | Machine Translation Evaluation |
| Dataset (including any stated split) | WMT 21 |
| Metric | Score (en-ha) |
| Unit | PERCENT |
| Direction | Higher is better |
| Results loaded | 6 of 6 |
| 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

- [Score (en-ha)](<https://www.sota2.com/research/sota/machine-translation-evaluation-on-wmt-21.md?metric=Score%20(en-ha)>): Higher is better; unit: PERCENT.

- [Score (xh-zu)](<https://www.sota2.com/research/sota/machine-translation-evaluation-on-wmt-21.md?metric=Score%20(xh-zu)>): Higher is better; unit: PERCENT.

## Recorded results

| Method | Score (en-ha) | Result date | Report paper |
| --- | --- | --- | --- |
| Llama 4 Scout | 54.3 | Not recorded | 69b0f7292d603e6de3e45f20 |
| Llama 3.3 70B | 49 | Not recorded | 69b0f7292d603e6de3e45f20 |
| Llama 3.3 70B | 43.2 | Not recorded | 69b0f7292d603e6de3e45f20 |
| Qwen 3 30B | 33.4 | Not recorded | 69b0f7292d603e6de3e45f20 |
| Llama 4 Scout | 30.3 | Not recorded | 69b0f7292d603e6de3e45f20 |
| Qwen 3 30B | 3.2 | Not recorded | 69b0f7292d603e6de3e45f20 |

### Llama 4 Scout

Score (en-ha): 54.3. Result ID: 69b12ae56dbb7075b95b9db1.

[Reporting paper](<https://www.sota2.com/research/paper/69b0f7292d603e6de3e45f20.md>)

[Method origin paper](<https://www.sota2.com/research/paper/69b0f7292d603e6de3e45f20.md>)

Recorded evaluation configuration: {"Shot":"Few"}. This record does not assert independent reproduction.

### Llama 3.3 70B

Score (en-ha): 49. Result ID: 69b12ae56dbb7075b95b9db3.

[Reporting paper](<https://www.sota2.com/research/paper/69b0f7292d603e6de3e45f20.md>)

[Method origin paper](<https://www.sota2.com/research/paper/69b0f7292d603e6de3e45f20.md>)

Recorded evaluation configuration: {"Shot":"Few"}. This record does not assert independent reproduction.

### Llama 3.3 70B

Score (en-ha): 43.2. Result ID: 69b12ae56dbb7075b95b9db4.

[Reporting paper](<https://www.sota2.com/research/paper/69b0f7292d603e6de3e45f20.md>)

[Method origin paper](<https://www.sota2.com/research/paper/69b0f7292d603e6de3e45f20.md>)

Recorded evaluation configuration: {"Shot":"Zero"}. This record does not assert independent reproduction.

### Qwen 3 30B

Score (en-ha): 33.4. Result ID: 69b12ae56dbb7075b95b9db5.

[Reporting paper](<https://www.sota2.com/research/paper/69b0f7292d603e6de3e45f20.md>)

[Method origin paper](<https://www.sota2.com/research/paper/69b0f7292d603e6de3e45f20.md>)

Recorded evaluation configuration: {"Shot":"Few"}. This record does not assert independent reproduction.

### Llama 4 Scout

Score (en-ha): 30.3. Result ID: 69b12ae56dbb7075b95b9db2.

[Reporting paper](<https://www.sota2.com/research/paper/69b0f7292d603e6de3e45f20.md>)

[Method origin paper](<https://www.sota2.com/research/paper/69b0f7292d603e6de3e45f20.md>)

Recorded evaluation configuration: {"Shot":"Zero"}. This record does not assert independent reproduction.

### Qwen 3 30B

Score (en-ha): 3.2. Result ID: 69b12ae56dbb7075b95b9db6.

[Reporting paper](<https://www.sota2.com/research/paper/69b0f7292d603e6de3e45f20.md>)

[Method origin paper](<https://www.sota2.com/research/paper/69b0f7292d603e6de3e45f20.md>)

Recorded evaluation configuration: {"Shot":"Zero"}. This record does not assert independent reproduction.

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