Llmfit
lastmod 2026-03-03created 2026-03-031 edits
A terminal tool that finds LLM models that fit your hardware and optimizes them automatically
About
- A tool that, with a single command, finds which of hundreds of LLM models and providers can actually run on your system's RAM, CPU, and GPU
- Scores each model on quality, speed, fit, and context and shows whether it can run; supports both TUI (terminal UI) and CLI modes
- Integrates multi-GPU, MoE architectures, dynamic quantization, speed estimation, and local runtimes (Ollama, llama.cpp, MLX)
- Analyzes the run mode (GPU, CPU+GPU, CPU) and fit level (Perfect, Good, Marginal, Too Tight) for each model and suggests the optimal combination
- Provides hardware-based automated model selection for developers who want to use LLMs efficiently in a local environment