▚ Awesome list · CC0

Awesome CPU-First AI

Training needs GPUs. Inference usually doesn't.

A curated, evidence-backed list of runtimes, quantization formats, benchmarks, and cost math for running AI and LLM inference on the CPU you already have — servers, laptops, edge devices, and serverless functions. Start with CPU; justify the GPU.

01Most inference never needed a GPU

Independent analysis of the Hugging Face ecosystem shows the models people actually run are small — small enough to fit comfortably on a modern CPU once quantized.

92%
of HF downloads go to models under 1B params
406M
median downloaded model size
~60%
of models are ≤ 13B params (quantized)
$5,740
/yr saved: 7B Q4 on CPU vs idle GPU

02When to choose CPU vs GPU for inference

This list is GPU-skeptical, not GPU-hostile. Here's the honest split — the full decision flowchart and "when you actually want a GPU" section live in the repo.

DimensionLean CPULean GPU
WorkloadInferenceTraining, fine-tuning
Model size≤ 13B (quantized)70B+ dense
ThroughputSingle-digit req/sHundreds req/s, batched
Latency SLA100 ms – 2 s TTFT ok< 50 ms TTFT
Context length≤ 8K tokens32K+ prefill-heavy
DeploymentEdge, serverless, SBCDedicated cluster
Cost / availabilityUbiquitous, no VRAM cap5–20× more, VRAM ceiling

03CPU-native runtimes & inference engines

Engines built for CPU as a primary target — not GPU frameworks that fall back to a slow CPU path.

llama.cpp

C/C++ LLM inference designed for CPU from day one; GGUF, x86/ARM/RISC-V.

ONNX Runtime (CPU EP)

Production-grade CPU execution provider with operator fusion and INT8.

OpenVINO

Intel's toolkit targeting x86 CPU as first-class hardware.

ncnn

Dependency-free mobile/embedded inference for ARM and x86.

ExecuTorch

PyTorch on-device runtime; ARM CPU (XNNPACK) is the primary target.

llamafile

Single-file LLM executables that run on CPU with no install.

See all runtimes, plus a format/arch/OS comparison table, in the repo.

04What's inside

Plus companion guides: a cost calculator, green-inference power math, benchmark methodology, and troubleshooting.

05Contribute

Know a CPU-native runtime, a real benchmark, or a broken link? Contributions are welcome — there are one-click issue templates for suggesting a resource, reporting a stale link, or disputing a benchmark figure.