WMT
Benchmarks
Task NameDataset NameSOTA ResultTrendResults
WMT'14 English-German (En-De) (test)
27.27BLEU
4
WMT EN-CS 14 (test)
28.2BLEU
4
WMT'14 EN-CS (val)
4.38Perplexity
4
WMT EN-FR 14 (val)
3.13Perplexity
4
WMT Zh-En
24.4BLEU
4
WMT11 English-Spanish (Sp->En) (test)
44P@1
4
WMT11 English-Spanish (En->Sp) (test)
39P@1
4
WMT En-Zh 2019 (newstest)
38.2SacreBLEU
4
WMT En-De 2020 (test)
34.4SacreBLEU
4
WMT14
2.72Average Acceptance Length (tau)
3
WMT en→xx 24++
88.89Score
3
WMT German-to-English 14 (test)
64.62BA Score
3
WMT German 3-shot 16 (test)
38.033BLEU (linewise)
3
WMT Metrics Shared Task
86.4SPA
3
WMT24pp
2.7N-Germanic
3
WMT En-De 19
35.2BLEU
3
WMT 22 (test)
0.92MCC
3
WMT 21 (test)
0.51MCC
3
WMT16 (test)
11.9FPR@95
3
WMT English-German 2021
0.733BLEURT
3
WMT Chinese-English 2021
69.2BLEURT
3
WMT Reverse set R 2017-2019 (test)
31.8BLEU
3
WMT biomedical 2018 (test)
30.3BLEU
3
WMT TS 2022 (test)
33.87De-En Score
3
WMT From-English 2022 (test)
61.7En->Zh Score
3