← All benchmarks

FLORES+ devtest (chrF++)

Chat & Writing

FLORES+ is a multi-way parallel translation benchmark built from English Wikimedia articles professionally translated into about 200 languages. This score uses the public devtest split of 1,012 sentences per language, translated sentence by sentence and scored with chrF++, and averages the out-of-English and into-English directions over the evaluated languages.

Top models (higher is better)