NLP for Arabic

Measured on OPUS, spaCy, Stanza, Tatoeba, NLLB-200, FLORES-200, Belebele.

Observatories covering Arabic

7 observatories measure NLP. Cards linking straight to this selection have data for it.

3,108,587,245 sentence pairs

33 corpora

4 pipeline modules

of 6 possible

6 pipeline processors

of 9 possible

68,433 sentences

in this language

Yes in nllb-200

FLORES-200

Open index →

No data for this selection

Evaluation benchmark for ~200 languages — the de facto baseline for whether a language is evaluable at all.

Yes in belebele