List your company
ResearchFormerly WizWand

LLM Prefill throughput on LLM Prompt Length 256 (pp256)

5,063Prefill Throughput (tok/s)uzu
343.481,568.742,7944,019.26Jul 1, 2026
Updated 4mo ago

Evaluation results

MethodLinks
uzuModel=Qwen3 0.6B, Quantization=4-bit, Hardware=Apple M4 Pro2026.075,063
llama.cppModel=Qwen2 0.5B, Quant=Q4, Hardware=Apple M4 Pro2026.074,987
llama.cppModel=Qwen2 0.5B, Quant=Q8, Hardware=Apple M4 Pro2026.074,914
uzuModel=Qwen3 0.6B, Quantization=8-bit, Hardware=Apple M4 Pro2026.074,887
BaseRTModel=Qwen3 0.6B, Quantization=Q4, Hardware=Apple M4 Pro2026.074,496
BaseRTModel=Qwen2 0.5B, Quant=Q4, Hardware=Apple M4 Pro2026.074,496
BaseRTModel=Qwen3 0.6B, Quantization=4-bit, Hardware=Apple M4 Pro2026.074,496
BaseRTModel=Qwen3 0.6B, Quantization=Q8, Hardware=Apple M4 Pro2026.074,434
BaseRTModel=Qwen2 0.5B, Quant=Q8, Hardware=Apple M4 Pro2026.074,434
BaseRTModel=Qwen3 0.6B, Quantization=8-bit, Hardware=Apple M4 Pro2026.074,434
MLXModel=Qwen3 0.6B, Quantization=Q4, Hardware=Apple M4 Pro2026.074,263
MLXModel=Qwen3 0.6B, Quantization=Q8, Hardware=Apple M4 Pro2026.074,255
llama.cppModel=Llama 3.2 1B, Quant=Q4, Hardware=Apple M4 Pro2026.072,714
MLXModel=Llama 3.2 1B, Quantization=Q4, Hardware=Apple M4 Pro2026.072,698
llama.cppModel=Llama 3.2 1B, Quant=Q8, Hardware=Apple M4 Pro2026.072,673
MLXModel=Llama 3.2 1B, Quantization=Q8, Hardware=Apple M4 Pro2026.072,620
uzuModel=Llama 3.2 1B, Quantization=4-bit, Hardware=Apple M4 Pro2026.072,604
BaseRTModel=Llama 3.2 1B, Quantization=Q4, Hardware=Apple M4 Pro2026.072,577
BaseRTModel=Llama 3.2 1B, Quant=Q4, Hardware=Apple M4 Pro2026.072,577
BaseRTModel=Llama 3.2 1B, Quantization=4-bit, Hardware=Apple M4 Pro2026.072,577
uzuModel=Llama 3.2 1B, Quantization=8-bit, Hardware=Apple M4 Pro2026.072,545
BaseRTModel=Llama 3.2 1B, Quantization=Q8, Hardware=Apple M4 Pro2026.072,517
BaseRTModel=Llama 3.2 1B, Quant=Q8, Hardware=Apple M4 Pro2026.072,517
BaseRTModel=Llama 3.2 1B, Quantization=8-bit, Hardware=Apple M4 Pro2026.072,517
llama.cppModel=Gemma 2 2B, Quant=Q4, Hardware=Apple M4 Pro2026.071,281
BaseRTModel=Gemma 2 2B, Quant=Q4, Hardware=Apple M4 Pro2026.071,193
BaseRTModel=Gemma 2 2B, Quant=Q8, Hardware=Apple M4 Pro2026.071,167
llama.cppModel=Gemma 2 2B, Quant=Q8, Hardware=Apple M4 Pro2026.07976
llama.cppModel=Llama 3.2 3B, Quant=Q8, Hardware=Apple M4 Pro2026.07928
uzuModel=Llama 3.2 3B, Quantization=4-bit, Hardware=Apple M4 Pro2026.07923
llama.cppModel=Llama 3.2 3B, Quant=Q4, Hardware=Apple M4 Pro2026.07916
BaseRTModel=Llama 3.2 3B, Quantization=Q4, Hardware=Apple M4 Pro2026.07906
BaseRTModel=Llama 3.2 3B, Quant=Q4, Hardware=Apple M4 Pro2026.07906
BaseRTModel=Llama 3.2 3B, Quantization=4-bit, Hardware=Apple M4 Pro2026.07906
MLXModel=Llama 3.2 3B, Quantization=Q4, Hardware=Apple M4 Pro2026.07901
MLXModel=Llama 3.2 3B, Quantization=Q8, Hardware=Apple M4 Pro2026.07897
uzuModel=Llama 3.2 3B, Quantization=8-bit, Hardware=Apple M4 Pro2026.07895
BaseRTModel=Llama 3.2 3B, Quantization=Q8, Hardware=Apple M4 Pro2026.07888
BaseRTModel=Llama 3.2 3B, Quant=Q8, Hardware=Apple M4 Pro2026.07888
BaseRTModel=Llama 3.2 3B, Quantization=8-bit, Hardware=Apple M4 Pro2026.07888
BaseRTModel=Qwen3 30B-A3B, Quantization=Q4, Hardware=Apple M4 Pro2026.07809
BaseRTModel=Mistral 7B v0.3, Quant=Q4, Hardware=Apple M4 Pro2026.07809
BaseRTModel=Gemma 4 26B-A4B, Quantization=Q4, Hardware=Apple M4 Pro2026.07704
BaseRTModel=Llama 3.1 8B, Quant=Q4, Hardware=Apple M4 Pro2026.07704
MLXModel=Gemma 4 26B-A4B, Quantization=Q4, Hardware=Apple M4 Pro2026.07576
MLXModel=Qwen3 30B-A3B, Quantization=Q4, Hardware=Apple M4 Pro2026.07571
llama.cppModel=Mistral 7B v0.3, Quant=Q4, Hardware=Apple M4 Pro2026.07550
llama.cppModel=Llama 3.1 8B, Quant=Q4, Hardware=Apple M4 Pro2026.07525