🧠llmfit
Hundreds of models & providers. One command to find what runs on your hardware.
⭐ Very popular: 32k stars, gaining about 635 a week
View on GitHub ↗repo profile
momentum
durability
bus factor = how many people it takes to cover more than half the commits (6 months). 1 is a solo project; higher means the work is spread across a team. top-author share is the single busiest author's slice of those commits.
since we covered it
why it's a big deal
- Answers a common question for anyone running models locally, which specific LLMs will actually fit on their RAM, CPU, and GPU, without trial-and-error downloads.
- Ranks models by a composite of quality, speed, memory fit, and context, and estimates tokens per second, so users can pick before committing to a multi-gigabyte download.
- Covers hundreds of models across providers like Meta, Mistral, Qwen, Google, Microsoft, and DeepSeek, and ties into local runtimes including Ollama, llama.cpp, MLX, LM Studio, and Docker Model Runner.
under the hood
- Written in Rust, with a ratatui and crossterm terminal UI plus CLI subcommands like system, fit, recommend, plan, list, search, serve, and bench.
- Detects hardware through sysinfo for RAM and CPU and vendor tools such as nvidia-smi, rocm-smi, system_profiler, and npu-smi for GPUs and NPUs, aggregating across multiple GPUs.
- Ships a HuggingFace-sourced model database embedded at compile time, computes memory needs across quantization levels, detects Mixture-of-Experts models to count only active experts, and estimates speed from a memory-bandwidth-bound formula backed by a GPU lookup table.
Radar summary, generated from the project's public sources
star history
- PR#28 4k 2026-02-25
- now 32k + 28k since first covered
curve is sampled from GitHub's star history, plus our own daily readings since we covered it; the dashed stretch is before we first covered it, the solid line since. figures at coverage are the numbers we printed then (approx.), current count is live.
understory
Quietly building: more output than attention, for now.
- output, commits & releases
- clout, star velocity
output = commits & releases; clout = star velocity, both 0 to 100 monthly indices; the gap where output runs above clout is the understory. The understory →
covered in
-
Lightweight fine-tuning workflows for LLMs
similar projects
compare these →- 🧠 Jan
TypeScript
Jan is an open source alternative to ChatGPT that runs 100% offline on your computer.
44k ACTIVE - 🧩 TensorZero
leaner, 12k stars
TensorZero is an open-source LLMOps platform that unifies an LLM gateway, observability, evaluation, optimization, and experimentation.
12k OFF THE RADAR - 🧰 Yek
leaner, 2k stars
A fast Rust based tool to serialize text-based files in a repository or directory for LLM consumption
2k ACTIVE
comments
Sign in with GitHub to add your blip on llmfit.