Back to Catalog
AlexsJones GitHub avatar

llmfit

Terminal tool to find LLM models that run well on your hardware

Demo

Terminal preview
llmfit

Overview

llmfit is a Rust terminal interface for right-sizing LLM models to system hardware.

It detects RAM, CPU cores, and GPU VRAM, calculates composite fit scores, estimates tokens-per-second throughput, and compares model memory requirements.

Hundreds of models & providers. One command to find what runs on your hardware.

Keyboard Shortcuts

Navigate through model catalog rowsj / k or ↓ / ↑
Search and filter models by name/
Cycle fit filter: All, Runnable, Perfect, Good, or Marginalf
Cycle sort column: Score, Params, Mem%, Ctx, Date, or Use Cases
Open Hardware Simulation popup for overriding VRAM, RAM, or CPUS
Filter by Provider, Use Case, Capability, License, or Runtime/backendP / U / C / L / R
Visual-range multi-select (v) or select mode for column-header actions (V)v / V
Toggle help overlay or dismiss active modal windowh / Esc

Latest Release

v1.1.15
Released: Sep 10, 2026

1.1.15 (2026-09-10)

Bug Fixes

  • bench: identify Ferrum and vLLM by endpoint owner (#994) (b3e09fd)
  • bench: normalize latency formatting (#1001) (1e7bdb3)
  • tui: render only visible model rows (#1017) (ff7a70b)

What's Changed

New Contributors

Full Changelog: https://github.com/AlexsJones/llmfit/compare/v1.1.14...v1.1.15

Related TUIs

Discussions

0 comments
Loading comments...
Alvin By Alvin