| ID | Language | Benchmark | Output | bleu |
|---|---|---|---|---|
| 0 | rus-eng | flores101-devtest | show | 37.3 |
| 1 | rus-eng | flores200-devtest | show | 37.3 |
| 2 | rus-eng | newstest2012 | show | 40 |
| 3 | rus-eng | newstest2013 | show | 38.3 |
| 4 | rus-eng | newstest2014 | show | 42.2 |
| 5 | rus-eng | newstest2015 | show | 45 |
| 6 | rus-eng | newstest2016 | show | 40 |
| 7 | rus-eng | newstest2017 | show | 51.9 |
| 8 | rus-eng | newstest2018 | show | 38.6 |
| 9 | rus-eng | newstest2019 | show | 39 |
| 10 | rus-eng | newstest2020 | show | 38 |
| 11 | rus-eng | newstestB2020 | show | 40.1 |
| 12 | rus-eng | tatoeba-test-v2020-07-28 | show | 57.1 |
| 13 | rus-eng | tatoeba-test-v2021-03-30 | show | 56.1 |
| 14 | rus-eng | tatoeba-test-v2021-08-07 | show | 55 |
| 15 | rus-eng | tico19-test | show | 35.5 |
| average | 43.212 |