Our new X account is live! Follow @wizwand_team for updates
Home
/
Benchmarks
Massive Multitask Language Understanding on MMLU (Performance Profile)
Loading...
56.6
MMLU
Qwen3-4B + FBS-Full (ours)
54.624
55.137
55.65
56.163
Jan 29, 2026
MMLU
Latency (ms)
TFLOPs (rel)
Bypass/Skip-Ratio (%)
Updated 4d ago
Evaluation Results
Method
Method
Links
MMLU
Latency (ms)
TFLOPs (rel)
Bypass/Skip-Ratio (%)
Qwen3-4B + FBS-Full (ours)
Backbone=Qwen3-4B
2026.01
56.6
532
0.7
36
Qwen3-4B + FBS-S1
Backbone=Qwen3-4B
2026.01
56.4
755
1.03
0
Qwen3-4B-Instruct (Baseline)
Backbone=Qwen3-4B
2026.01
55.1
760
1
0
Qwen3-4B + EAGLE-2 (Group A)
Backbone=Qwen3-4B
2026.01
55
555
0.74
30
Qwen3-4B + Lookahead (Group A)
Backbone=Qwen3-4B
2026.01
55
595
0.82
15
Qwen3-4B + SpecDec (Group A)
Backbone=Qwen3-4B
2026.01
54.9
646
0.9
22
Qwen3-4B + Medusa (Group A)
Backbone=Qwen3-4B
2026.01
54.7
570
0.8
18
Feedback
Search any
task
Search any
task