MobileMoE: Scaling On-Device Mixture of Experts

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Chen, Yanbei, Huang, Hanxian, Chang, Ernie, Szwejbka, Jacob, Desai, Digant, Liu, Zechun, Chandra, Vikas, Krishnamoorthi, Raghuraman
Format: Preprint
Published: 2026
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866911721481830400
author Chen, Yanbei
Huang, Hanxian
Chang, Ernie
Szwejbka, Jacob
Desai, Digant
Liu, Zechun
Chandra, Vikas
Krishnamoorthi, Raghuraman
author_facet Chen, Yanbei
Huang, Hanxian
Chang, Ernie
Szwejbka, Jacob
Desai, Digant
Liu, Zechun
Chandra, Vikas
Krishnamoorthi, Raghuraman
contents Mixture-of-Experts (MoE) has become the de facto architecture for hundred-billion-parameter language models, yet its advantages at sub-billion scales for on-device deployment remain largely unexplored. To close this gap, we present MobileMoE, a family of on-device MoE language models with sub-billion active parameters (0.3-0.9B active and 1.3-5.3B total) that establish a new Pareto frontier for on-device LLMs. We first formulate an on-device MoE scaling law that jointly optimizes MoE architecture under mobile memory and compute constraints, identifying an on-device sweet spot - moderate sparsity with fine-grained and shared experts - that is simultaneously memory and compute-optimal. Building on the derived architectures, we train MobileMoE with a four-stage recipe covering pre-training, mid-training, instruction fine-tuning, and quantization-aware training, all on open-source datasets. Across 14 benchmarks, MobileMoE matches or exceeds leading on-device dense LLMs with 2-4$\times$ fewer inference FLOPs, and matches or surpasses the state-of-the-art MoE OLMoE-1B-7B with up to 60% fewer parameters. To bridge the last mile to mobile deployment, we provide the first efficient MoE inference on commodity smartphones with comprehensive on-device profiling. At comparable INT4 weight memory, MobileMoE-S delivers $1.8$-$3.8\times$ faster prefill and $2.2$-$3.4\times$ faster decode than the dense baseline MobileLLM-Pro.
format Preprint
id arxiv_https___arxiv_org_abs_2605_27358
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle MobileMoE: Scaling On-Device Mixture of Experts
Chen, Yanbei
Huang, Hanxian
Chang, Ernie
Szwejbka, Jacob
Desai, Digant
Liu, Zechun
Chandra, Vikas
Krishnamoorthi, Raghuraman
Machine Learning
Artificial Intelligence
Computation and Language
Mixture-of-Experts (MoE) has become the de facto architecture for hundred-billion-parameter language models, yet its advantages at sub-billion scales for on-device deployment remain largely unexplored. To close this gap, we present MobileMoE, a family of on-device MoE language models with sub-billion active parameters (0.3-0.9B active and 1.3-5.3B total) that establish a new Pareto frontier for on-device LLMs. We first formulate an on-device MoE scaling law that jointly optimizes MoE architecture under mobile memory and compute constraints, identifying an on-device sweet spot - moderate sparsity with fine-grained and shared experts - that is simultaneously memory and compute-optimal. Building on the derived architectures, we train MobileMoE with a four-stage recipe covering pre-training, mid-training, instruction fine-tuning, and quantization-aware training, all on open-source datasets. Across 14 benchmarks, MobileMoE matches or exceeds leading on-device dense LLMs with 2-4$\times$ fewer inference FLOPs, and matches or surpasses the state-of-the-art MoE OLMoE-1B-7B with up to 60% fewer parameters. To bridge the last mile to mobile deployment, we provide the first efficient MoE inference on commodity smartphones with comprehensive on-device profiling. At comparable INT4 weight memory, MobileMoE-S delivers $1.8$-$3.8\times$ faster prefill and $2.2$-$3.4\times$ faster decode than the dense baseline MobileLLM-Pro.
title MobileMoE: Scaling On-Device Mixture of Experts
topic Machine Learning
Artificial Intelligence
Computation and Language
url https://arxiv.org/abs/2605.27358