ABEJA-Qwen2.5-32B-Japanese-v1.0: benchmark results

Provider: Other. Access: Open.

Unified ELO 1547 ± 1, rank #452 of 1392 rated models, from 80 benchmark results.

Strongest benchmark results

BenchmarkScoreMetricPercentile
Open Japanese LLM - Jcommonsenseqa Exact Match97.14Score (%)99.8
Open Japanese LLM - Jmmlu Exact Match77.69Score (%)99.3
Open Japanese LLM - MC91.7Score (%)98.3
Open Japanese LLM - HE78.81Score (%)98.1
Open Japanese LLM - Kuci Exact Match85.29Score (%)97.1
Open Japanese LLM - Wiki PAS SET F111.64Score (%)96
Open Japanese LLM - Jemhopqa Char F165.36Score (%)95.9
Open Japanese LLM - NLI81.46Score (%)95
Open Japanese LLM - Commonsensemoralja Exact Match92.69Score (%)94.2
Open Japanese LLM - Jsts Pearson90.42Score (%)92.5
Open Japanese LLM - Jnli Exact Match89.52Score (%)92.3
Open Japanese LLM - MR94Score (%)91.3

Interactive version: theaggregate.ai/model?slug=abeja-qwen2-5-32b-japanese-v1-0 · How It Works · Data refreshed daily, snapshot 2026-09-05.