Share your thoughts, 1 month free Claude Pro on us
See more
Home
/
Benchmarks
Decode Throughput on Qwen3 30B-A3B
Loading...
84.1
Throughput (tokens/s)
BaseRT
80.564
81.482
82.4
83.318
Jul 1, 2026
Throughput (tokens/s)
Speedup Ratio vs. llama.cpp
Speedup Ratio vs. MLX
Updated 24d ago
Evaluation Results
Method
Method
Links
Throughput (tokens/s)
Speedup Ratio vs. llama.cpp
Speedup Ratio vs. MLX
BaseRT
Quantization=Q4, Hardw...
2026.07
84.1
1.04
1.01
MLX
Quantization=Q4, Hardw...
2026.07
83.1
-
-
llama.cpp
Quantization=Q4, Hardw...
2026.07
80.7
-
-
Feedback
Search any
task
Search any
task
