List your company
ResearchFormerly WizWand

LLM Prefill throughput on LLM Prompt Length 512 (pp512)

5,128Throughput (tok/s)uzu
428.241,648.372,868.54,088.63Jul 1, 2026
Updated 4mo ago

Evaluation results

MethodLinks
uzuModel=Qwen3 0.6B, Quantization=4-bit, Hardware=Apple M4 Pro2026.075,128
uzuModel=Qwen3 0.6B, Quantization=8-bit, Hardware=Apple M4 Pro2026.075,001
llama.cppModel=Qwen2 0.5B, Quant=Q4, Hardware=Apple M4 Pro2026.074,993
llama.cppModel=Qwen2 0.5B, Quant=Q8, Hardware=Apple M4 Pro2026.074,933
BaseRTModel=Qwen3 0.6B, Quantization=Q4, Hardware=Apple M4 Pro2026.074,690
BaseRTModel=Qwen2 0.5B, Quant=Q4, Hardware=Apple M4 Pro2026.074,690
BaseRTModel=Qwen3 0.6B, Quantization=4-bit, Hardware=Apple M4 Pro2026.074,690
MLXModel=Qwen3 0.6B, Quantization=Q4, Hardware=Apple M4 Pro2026.074,658
MLXModel=Qwen3 0.6B, Quantization=Q8, Hardware=Apple M4 Pro2026.074,654
BaseRTModel=Qwen3 0.6B, Quantization=Q8, Hardware=Apple M4 Pro2026.074,623
BaseRTModel=Qwen2 0.5B, Quant=Q8, Hardware=Apple M4 Pro2026.074,623
BaseRTModel=Qwen3 0.6B, Quantization=8-bit, Hardware=Apple M4 Pro2026.074,623
MLXModel=Llama 3.2 1B, Quantization=Q4, Hardware=Apple M4 Pro2026.072,817
MLXModel=Llama 3.2 1B, Quantization=Q8, Hardware=Apple M4 Pro2026.072,766
llama.cppModel=Llama 3.2 1B, Quant=Q4, Hardware=Apple M4 Pro2026.072,728
llama.cppModel=Llama 3.2 1B, Quant=Q8, Hardware=Apple M4 Pro2026.072,702
uzuModel=Llama 3.2 1B, Quantization=4-bit, Hardware=Apple M4 Pro2026.072,642
BaseRTModel=Llama 3.2 1B, Quantization=Q4, Hardware=Apple M4 Pro2026.072,615
BaseRTModel=Llama 3.2 1B, Quant=Q4, Hardware=Apple M4 Pro2026.072,615
BaseRTModel=Llama 3.2 1B, Quantization=4-bit, Hardware=Apple M4 Pro2026.072,615
uzuModel=Llama 3.2 1B, Quantization=8-bit, Hardware=Apple M4 Pro2026.072,597
BaseRTModel=Llama 3.2 1B, Quantization=Q8, Hardware=Apple M4 Pro2026.072,564
BaseRTModel=Llama 3.2 1B, Quant=Q8, Hardware=Apple M4 Pro2026.072,564
BaseRTModel=Llama 3.2 1B, Quantization=8-bit, Hardware=Apple M4 Pro2026.072,564
llama.cppModel=Gemma 2 2B, Quant=Q4, Hardware=Apple M4 Pro2026.071,292
BaseRTModel=Gemma 2 2B, Quant=Q4, Hardware=Apple M4 Pro2026.071,208
BaseRTModel=Gemma 2 2B, Quant=Q8, Hardware=Apple M4 Pro2026.071,187
llama.cppModel=Gemma 2 2B, Quant=Q8, Hardware=Apple M4 Pro2026.07983
MLXModel=Llama 3.2 3B, Quantization=Q8, Hardware=Apple M4 Pro2026.07943
llama.cppModel=Llama 3.2 3B, Quant=Q8, Hardware=Apple M4 Pro2026.07932
uzuModel=Llama 3.2 3B, Quantization=4-bit, Hardware=Apple M4 Pro2026.07929
MLXModel=Llama 3.2 3B, Quantization=Q4, Hardware=Apple M4 Pro2026.07919
uzuModel=Llama 3.2 3B, Quantization=8-bit, Hardware=Apple M4 Pro2026.07905
BaseRTModel=Llama 3.2 3B, Quantization=Q4, Hardware=Apple M4 Pro2026.07899
BaseRTModel=Llama 3.2 3B, Quant=Q4, Hardware=Apple M4 Pro2026.07899
BaseRTModel=Llama 3.2 3B, Quantization=4-bit, Hardware=Apple M4 Pro2026.07899
BaseRTModel=Llama 3.2 3B, Quantization=Q8, Hardware=Apple M4 Pro2026.07893
BaseRTModel=Llama 3.2 3B, Quant=Q8, Hardware=Apple M4 Pro2026.07893
BaseRTModel=Llama 3.2 3B, Quantization=8-bit, Hardware=Apple M4 Pro2026.07893
llama.cppModel=Llama 3.2 3B, Quant=Q4, Hardware=Apple M4 Pro2026.07871
BaseRTModel=Qwen3 30B-A3B, Quantization=Q4, Hardware=Apple M4 Pro2026.07838
BaseRTModel=Mistral 7B v0.3, Quant=Q4, Hardware=Apple M4 Pro2026.07838
BaseRTModel=Gemma 4 26B-A4B, Quantization=Q4, Hardware=Apple M4 Pro2026.07721
BaseRTModel=Llama 3.1 8B, Quant=Q4, Hardware=Apple M4 Pro2026.07721
llama.cppModel=Mistral 7B v0.3, Quant=Q4, Hardware=Apple M4 Pro2026.07668
MLXModel=Qwen3 30B-A3B, Quantization=Q4, Hardware=Apple M4 Pro2026.07664
MLXModel=Gemma 4 26B-A4B, Quantization=Q4, Hardware=Apple M4 Pro2026.07639
llama.cppModel=Llama 3.1 8B, Quant=Q4, Hardware=Apple M4 Pro2026.07609