List your company

Inference Energy Consumption on Qwen3-MoE 30B (inference)

3.8Energy per Run (J)NPUMoE
3.44525.84018.23510.6299Apr 20, 2026

Evaluation results

MethodLinks
NPUMoEMode=Ours, Hardware=M2 Ultra2026.043.81.231.21.371
anemllMode=anemll, Hardware=M2 Ultra2026.0412.138.870.262.993.19
coreml-onlyMode=coreml-only, Hardware=M2 Ultra2026.0412.679.6603.013.33