Share your thoughts, 1 month free Claude Pro on us
See more
Home
/
Benchmarks
Code Generation on HumanEval (Speedup)
Loading...
4.06
Speedup
FailFast
1.2936
2.0118
2.73
3.4482
Dec 23, 2025
Speedup
Updated 3mo ago
Evaluation Results
Method
Method
Links
Speedup
FailFast
Target Model=Qwen2.5-3...
2025.12
4.06
FailFast
Target Model=Qwen2.5-1...
2025.12
3.41
EAGLE-3 (w/ draft tree)
Target Model=Qwen2.5-3...
2025.12
3.31
Fast-dLLM
Target Model=Qwen2.5-3...
2025.12
3.16
AR Draft Model
Target Model=Qwen2.5-3...
2025.12
2.72
FailFast
Target Model=Qwen2.5-7...
2025.12
2.71
EAGLE-3 (w/ draft tree)
Target Model=Qwen2.5-1...
2025.12
2.68
EAGLE-3 (w/ draft tree)
Target Model=Qwen2.5-7...
2025.12
2.39
Fast-dLLM
Target Model=Qwen2.5-1...
2025.12
2.23
AR Draft Model
Target Model=Qwen2.5-1...
2025.12
1.91
Fast-dLLM
Target Model=Qwen2.5-7...
2025.12
1.61
AR Draft Model
Target Model=Qwen2.5-7...
2025.12
1.4
Feedback
Search any
task
Search any
task