Benchmark

BEAR-Bench

What is BEAR-Bench?

BEAR-Bench evaluates multimodal models on bilingual English-Russian text-dense business and scientific documents with 1,000 human-annotated questions.

Released: 2026-08-18
Evaluates: Bilingual reasoning, Finance, Multimodal, Multimodal reasoning, OCR-grounded reasoning, Professional document reasoning, Reasoning, Scientific Research & AI for Science
Openness: unknown
Importer: Claire Radar
Review status: source-reviewed
Reported scores: 0

Source provenance

Original evidence https://arxiv.org/abs/2608.17895

BEAR-Bench paper

Paper https://arxiv.org/abs/2608.17895

No reported scores are on record for this benchmark yet.

Which sources cite BEAR-Bench?

BEAR-Bench

Related Bilingual reasoning, Finance, Multimodal, Multimodal reasoning, OCR-grounded reasoning, Professional document reasoning, Reasoning, Scientific Research & AI for Science benchmarks

3DCodeBench
A-Bench
A-OKVQA
ActionAtlas
ActiveFly-Bench
AdvancedMathBench
AeroGround
AFDBench