Our new X account is live! Follow @wizwand_team for updates
Home
/
Benchmarks
Chinese Massive Multitask Language Understanding on CMMLU (Performance and Efficiency)
Loading...
57.4
CMMLU Score
Qwen3-4B + FBS-Full (ours)
55.632
56.091
56.55
57.009
Jan 29, 2026
CMMLU Score
Latency (ms)
TFLOPs (rel)
Bypass/Skip Ratio (%)
Updated 4d ago
Evaluation Results
Method
Method
Links
CMMLU Score
Latency (ms)
TFLOPs (rel)
Bypass/Skip Ratio (%)
Qwen3-4B + FBS-Full (ours)
Backbone=Qwen3-4B
2026.01
57.4
532
0.7
36
Qwen3-4B-Instruct (Baseline)
Backbone=Qwen3-4B
2026.01
55.7
760
1
0
Feedback
Search any
task
Search any
task