Benchmark collection
ReLE Chinese LLM Benchmark
Compare commercial and open models on a continuously updated Chinese-language evaluation.
Explore the original evidence
Explore the original evaluation, its methods, and reported findings.
Open evalCoverage
More than 370 commercial and open models in October 2026, scored on Chinese-language tasks across several ability areas.
Methodology
The maintainers run models on their own Chinese question sets and publish category and overall rankings in the repository.