Atlas

Benchmarks

← All benchmarks

Belebele

General QA

Belebele is a fully parallel multiple-choice reading-comprehension benchmark in 122 language variants. Each question is grounded in a short FLORES passage and offers four options, and the score is mean accuracy across the language variants.

Top models (higher is better)

ModelScore
Muse Glimmer 30B87.8
Loading Atlas data…