List your company
ResearchFormerly WizWand

LLM Prefill throughput on LLM Prompt Length 2048 (pp2048)

4,447Prefill Throughput (tok/s)MLX
402.441,452.472,502.53,552.53Jul 1, 2026
Updated 4mo ago

Evaluation results

MethodLinks
MLXModel=Qwen3 0.6B, Quantization=Q8, Hardware=Apple M4 Pro2026.074,447
MLXModel=Qwen3 0.6B, Quantization=Q4, Hardware=Apple M4 Pro2026.074,439
llama.cppModel=Qwen2 0.5B, Quant=Q4, Hardware=Apple M4 Pro2026.074,075
llama.cppModel=Qwen2 0.5B, Quant=Q8, Hardware=Apple M4 Pro2026.074,045
BaseRTModel=Qwen3 0.6B, Quantization=Q4, Hardware=Apple M4 Pro2026.073,680
BaseRTModel=Qwen2 0.5B, Quant=Q4, Hardware=Apple M4 Pro2026.073,680
BaseRTModel=Qwen3 0.6B, Quantization=Q8, Hardware=Apple M4 Pro2026.073,648
BaseRTModel=Qwen2 0.5B, Quant=Q8, Hardware=Apple M4 Pro2026.073,648
MLXModel=Llama 3.2 1B, Quantization=Q8, Hardware=Apple M4 Pro2026.072,783
MLXModel=Llama 3.2 1B, Quantization=Q4, Hardware=Apple M4 Pro2026.072,714
llama.cppModel=Llama 3.2 1B, Quant=Q8, Hardware=Apple M4 Pro2026.072,520
llama.cppModel=Llama 3.2 1B, Quant=Q4, Hardware=Apple M4 Pro2026.072,489
BaseRTModel=Llama 3.2 1B, Quantization=Q4, Hardware=Apple M4 Pro2026.072,453
BaseRTModel=Llama 3.2 1B, Quant=Q4, Hardware=Apple M4 Pro2026.072,453
BaseRTModel=Llama 3.2 1B, Quantization=Q8, Hardware=Apple M4 Pro2026.072,418
BaseRTModel=Llama 3.2 1B, Quant=Q8, Hardware=Apple M4 Pro2026.072,418
llama.cppModel=Gemma 2 2B, Quant=Q4, Hardware=Apple M4 Pro2026.071,209
BaseRTModel=Gemma 2 2B, Quant=Q4, Hardware=Apple M4 Pro2026.071,169
BaseRTModel=Gemma 2 2B, Quant=Q8, Hardware=Apple M4 Pro2026.071,153
llama.cppModel=Gemma 2 2B, Quant=Q8, Hardware=Apple M4 Pro2026.07971
MLXModel=Llama 3.2 3B, Quantization=Q4, Hardware=Apple M4 Pro2026.07934
MLXModel=Llama 3.2 3B, Quantization=Q8, Hardware=Apple M4 Pro2026.07900
llama.cppModel=Llama 3.2 3B, Quant=Q4, Hardware=Apple M4 Pro2026.07864
llama.cppModel=Llama 3.2 3B, Quant=Q8, Hardware=Apple M4 Pro2026.07840
BaseRTModel=Llama 3.2 3B, Quantization=Q4, Hardware=Apple M4 Pro2026.07835
BaseRTModel=Llama 3.2 3B, Quant=Q4, Hardware=Apple M4 Pro2026.07835
BaseRTModel=Llama 3.2 3B, Quantization=Q8, Hardware=Apple M4 Pro2026.07792
BaseRTModel=Llama 3.2 3B, Quant=Q8, Hardware=Apple M4 Pro2026.07792
BaseRTModel=Qwen3 30B-A3B, Quantization=Q4, Hardware=Apple M4 Pro2026.07721
BaseRTModel=Mistral 7B v0.3, Quant=Q4, Hardware=Apple M4 Pro2026.07721
MLXModel=Qwen3 30B-A3B, Quantization=Q4, Hardware=Apple M4 Pro2026.07720
MLXModel=Gemma 4 26B-A4B, Quantization=Q4, Hardware=Apple M4 Pro2026.07644
BaseRTModel=Gemma 4 26B-A4B, Quantization=Q4, Hardware=Apple M4 Pro2026.07626
BaseRTModel=Llama 3.1 8B, Quant=Q4, Hardware=Apple M4 Pro2026.07626
llama.cppModel=Mistral 7B v0.3, Quant=Q4, Hardware=Apple M4 Pro2026.07601
llama.cppModel=Llama 3.1 8B, Quant=Q4, Hardware=Apple M4 Pro2026.07558