List your company

Chinese Massive Multitask Language Understanding on CMMLU (Performance and Efficiency)

57.4CMMLU ScoreQwen3-4B + FBS-Full (ours)
55.63256.09156.5557.009Jan 29, 2026

Evaluation results

MethodLinks
Qwen3-4B + FBS-Full (ours)Backbone=Qwen3-4B2026.0157.45320.736
Qwen3-4B-Instruct (Baseline)Backbone=Qwen3-4B2026.0155.776010