Atlas

Benchmarks

← All benchmarks

WMT24++ (chrF++)

Chat & Writing

WMT24++ extends the WMT24 test set to 55 languages and dialects across literary, news, social and speech domains, with new human references and post-edits. Each segment is translated in one turn and scored with chrF++ against the human references. This score averages the out-of-English and into-English directions over the 55 language pairs.

Top models (higher is better)

ModelScore
Muse Glimmer 30B53.4
Loading Atlas data…