moeshift
Memory efficient and high performance Mixture-of-Experts inference for Apple Silicon
What it does
Core capabilities at a glance
- Apple Silicon
- Llama CPP
- LLM Inference
- Local LLM
- Memory Optimization
- Mixture OF Experts
- MOE
Deep dive
The full breakdown - performance, comparisons, and setup
moeshiftLink to this section
moeshift is a local inference server - Memory efficient and high performance Mixture-of-Experts inference for Apple Silicon.
OverviewLink to this section
125 billion parameters. Up to 11 tokens per second. Under 10 GB of peak process memory. MoEShift is optimized for Apple Silicon: on a Mac mini M4 Pro with 24 GB of unified memory, it runs Qwen 3.8 Flash Next with roughly half the peak process memory of Strata and processes long prompts 2.6× faster.
MoEShift leaves room for your other apps, helping your Mac stay responsive during inference instead of grinding to a halt under memory pressure. Keep browsing, writing and coding while the model works—all locally, with no cloud subscription.
Qwen 3.8 Flash Next running locally with MoEShift, connected to OpenCode.
Alpha release. The core runtime is working, and the project is evolving quickly. You may encounter rough edges; your feedback helps shape the next release. Tested models and quantizations are listed below.
Both use the shared SSD expert runtime on Apple Silicon. See supported architectures.
Compared with Strata, a popular choice for local MoE inference, on a Mac mini M4 Pro with 24 GB unified memory. Both ran Qwen 3.8 Flash Next (125B), ISTA-DASLab GSQ-RCO IQ3_XXS, with an 8,192-token context, no cached prompt tokens and 256 output tokens. Strata: 0.2.7, build 0.1.52.
Time includes server-measured prompt processing and generation, excluding startup. Memory is macOS 'footprint -p' peak physical process footprint.
moeshift is open-source, written primarily in C++, with 113 GitHub stars under the MIT license. It was last updated on 2026-11-04.
Key capabilitiesLink to this section
From the project's documentation:
- A Mac with Apple Silicon (M-series).
- At least 10 GB of RAM.
- A fast SSD with space for both the GGUF and prepared expert files: at least
- Homebrew, the package manager for macOS. If you have never
InstallLink to this section
A quick way to get started (always check the official docs for the latest):
brew install davide-troiani/moeshift/moeshiftHow it fits a local-AI stackLink to this section
moeshift runs on your own hardware, so pair it with a model and a GPU sized to your needs. Use the VRAM calculator to pick a model that fits your card, and see what you can run for hardware guidance. Related local inference servers in the directory:
SourcesLink to this section
- Source code & docs: davide-troiani/moeshift
Stats from GitHub, 2026-11-13.
Frequently asked
Quick answers to common questions
What is moeshift?
moeshift is a inference-server tool for local AI workloads. Memory efficient and high performance Mixture-of-Experts inference for Apple Silicon
Is moeshift free and open source?
Yes, moeshift has 113 GitHub stars and is licensed under MIT. You can self-host it for free on macos.
What platforms does moeshift support?
moeshift runs on macos.
What hardware do I need for moeshift?
The hardware requirements depend on which models you run. Check our hardware directory for compatible GPUs and systems. moeshift has 113 GitHub stars and an active community.
Does moeshift support GPU acceleration?
moeshift supports GPU acceleration via CUDA, Metal, or Vulkan depending on your platform. For the best performance, pair it with an NVIDIA RTX 4090 or 5090.
What are the best alternatives to moeshift?
Popular alternatives include other inference-server tools in our directory. Browse our full collection at /tool for comparisons, community reviews, and benchmark data to find the right fit for your workflow.
How much does moeshift cost?
moeshift is free-open-source. It is completely free and open source to self-host.
Pairs well with
Complementary tools, models, and hardware