Share your thoughts, 1 month free Claude Pro on us
See more
Home
/
Benchmarks
Code Generation on HumanEval (TPS, Latency)
Loading...
28.24
TPS (Tokens/s)
MoE-SpAc
0.0664
7.3807
14.695
22.0093
Feb 12, 2026
TPS (Tokens/s)
Latency (ms)
Updated 1mo ago
Evaluation Results
Method
Method
Links
TPS (Tokens/s)
Latency (ms)
MoE-SpAc
2026.02
28.24
16.93
llama.cpp
speculative decoding=true
2026.02
18.78
24.01
llama.cpp
speculative decoding=f...
2026.02
14.74
30.57
HybriMoE
2026.02
11.96
38.18
Fate
2026.02
7.82
60.24
SP-MoE
2026.02
4.98
86.93
MoE-Infinity
2026.02
3.63
119.53
vLLM
2026.02
2.63
158.46
Mixtral Offload
2026.02
2.48
177.87
Accelerate
2026.02
1.15
393.6
Feedback
Search any
task
Search any
task