List your company

ResearchFormerly WizWand

Translation

Benchmarks

DatasetSOTA methodMetricTrendResultsLast updated

FLORES-200 it-en (devtest)

37.032sacreBLEU

35

Feb 26, 2026

FLORES-200 en-it (devtest)

33.831sacreBLEU

35

Feb 26, 2026

Teacher (70B SFT)

86.31COMET Score

34

Jun 23, 2026

lexical choice (val+test)

54.15BLEU

24

Mar 9, 2026

Konkani-Bench 200 samples (test)

gemini-flash-latest

86.76BLEU

22

Mar 26, 2026

Spec-Bench Trans.

6.41CR

21

Apr 15, 2026

41.8chrF

20

Jul 1, 2026

FLoRes+ All total

50.9ChrF++

16

Mar 18, 2026

FLoRes+ En-YY (total)

44.6ChrF++

16

Mar 18, 2026

FLoRes+ En-YY, v. low resource level

34.6ChrF++

16

Mar 18, 2026

FLoRes+ En-YY, low resource level

41ChrF++

16

Mar 18, 2026

FLoRes+ En-YY mid resource level

46.5ChrF++

16

Mar 18, 2026

FLoRes+ En-YY, high resource level

58.6ChrF++

16

Mar 18, 2026

FLoRes+ XX-En total

57.3ChrF++

16

Mar 18, 2026

FLoRes+ XX-En v. low resource level

52.3ChrF++

16

Mar 18, 2026

FLoRes+ XX-En, low resource level

54.6ChrF++

16

Mar 18, 2026

FLoRes+ XX-En mid resource level

58.3ChrF++

16

Mar 18, 2026

FLoRes+ XX-En, high resource level

65ChrF++

16

Mar 18, 2026

English-Spanish

76.7Accuracy

15

Feb 26, 2026

Sand-Glass Zh-to-En (human evaluation)

4.97Accuracy

14

Feb 26, 2026

Copyrighted Content without copyright notice 1.0 (test)

44.84Cosine Similarity

12

Feb 26, 2026

41.22ROUGE-L

12

Feb 26, 2026

Custom English-to-Hebrew

DictaLM-3.0-24B-Thinking

30.09Win Rate (GPT-4o Judge)

11

Feb 27, 2026

Supervised Model

39.9BLEU

9

Feb 26, 2026

OPUS English-Chinese (test)

41.4chrF

8

Jul 1, 2026