Repolode
Submit
AlexsJones

AlexsJones/llmfit

A Rust CLI and TUI that detects your CPU, RAM and GPU, then ranks open LLMs by how well they will run on your machine.

View on GitHub
Stars
38,129
Forks
2,456
Open issues
60
Last commit
yesterday

What is it?

llmfit answers a common question for people running local models: which open-source LLMs will fit on my hardware and run at a usable speed. It inspects CPU cores, system RAM, discrete and integrated GPUs, VRAM and unified memory, including NVIDIA CUDA, Apple Silicon, AMD ROCm and Intel OneAPI setups.

Each model in its catalog is scored across memory fit, estimated speed, quality and context length. The estimate accounts for parameter counts and quantization formats such as GGUF, AWQ, GPTQ and EXL2. Results are shown in an interactive terminal UI by default, or as a classic CLI table and JSON output for scripts.

The project also offers a web dashboard and a REST API served over HTTP, plus a benchmark command that measures tok/s against a running local provider such as Ollama, llama.cpp, MLX, Docker Model Runner or LM Studio.

Key features

  • 01Detects CPU, RAM, GPU, VRAM and unified memory across NVIDIA, Apple Silicon, AMD and Intel.
  • 02Ranks models by memory fit, estimated speed, quality and context length.
  • 03Offers an interactive TUI, a CLI with JSON output, and a web dashboard.
  • 04Exposes HTTP endpoints such as /api/v1/system and /api/v1/models.
  • 05Supports multi-GPU setups, MoE models and multiple quantization formats.

Use cases

  • Picking a local LLM before downloading multi-gigabyte model files.
  • Checking whether a laptop or workstation can run a model at an acceptable speed.
  • Feeding hardware-aware model recommendations into scripts or agents via JSON.
  • Sharing hardware benchmark results with the community to improve speed estimates.

Quick start

  1. 1

    Install with Homebrew (prebuilt binary)

    brew install AlexsJones/llmfit/llmfit
  2. 2

    Launch the interactive TUI

    llmfit

Things to know

  • Windows release binaries may be unsigned if the signing job fails, so verify signatures if you need a signed binary.
  • Speed and fit figures are estimates from specs unless you run benchmarks yourself.

Topics

#llm#local-llm#gguf#quantization#hardware#cli#tui#benchmark#localai#mlx

FAQ

What is llmfit?
A Rust CLI and TUI that detects your CPU, RAM and GPU, then ranks open LLMs by how well they will run on your machine. llmfit answers a common question for people running local models: which open-source LLMs will fit on my hardware and run at a usable speed. It inspects CPU cores, system RAM, discrete and integrated GPUs, VRAM and unified memory, including NVIDIA CUDA, Apple Silicon, AMD ROCm and Intel OneAPI setups.
Is llmfit free to use? What license does it have?
llmfit is open source under the MIT license. Check the license terms before using it in commercial products.
Is llmfit still maintained?
Actively maintained. The last commit to llmfit was on Oct 19, 2026.
What can I use llmfit for?
Picking a local LLM before downloading multi-gigabyte model files. Checking whether a laptop or workstation can run a model at an acceptable speed. Feeding hardware-aware model recommendations into scripts or agents via JSON. Sharing hardware benchmark results with the community to improve speed estimates.
How do I install llmfit?
Install with Homebrew (prebuilt binary): brew install AlexsJones/llmfit/llmfit Launch the interactive TUI: llmfit
Where can I get llmfit?
The source code and installation instructions are on GitHub: https://github.com/AlexsJones/llmfit

Summary generated by AI from the project's README. GitHub stats refresh daily.

Related projects