# Whichllm

> Find the best local LLM for your hardware using real benchmarks

- Canonical URL: https://www.fosshunter.com/tools/whichllm
- JSON export: https://www.fosshunter.com/tools/whichllm.json
- Website: https://github.com/Andyyyy64/whichllm
- Category: Developer Tools
- Open source: yes
- Self-hostable: yes
- Verified listing: no
- Featured: no
- Published at: 2026-06-12T05:14:01+00:00
- Catalog revision: `156355998`

## Repository metrics (point-in-time)

- Repository: https://github.com/Andyyyy64/whichllm
- Stars: 6818
- Forks: 376
- Open issues: 6
- License: MIT
- Last push: 2026-11-03T19:21:49+00:00
- Metrics observed at (synced_at): 2026-11-03T19:25:16+00:00
- Archived: no

Tags: LLM, AI, Developer Tools, Ollama, Silicon, Apple Silicon, CLI, Local LLM, Inference, Benchmarks
Tech stacks: CLI, GGUF, HuggingFace, python

## Features
- Detects local hardware capabilities including GPU and system VRAM. *(highlighted)*
- Provides live rankings from HuggingFace based on performance quality. *(highlighted)*
- Simulates performance on high-end GPUs before purchasing hardware. *(highlighted)*

## Description

Whichllm is a CLI tool that auto-detects hardware to rank top LLMs from HuggingFace. It uses recency-aware benchmarks to recommend models that perform best on your system, avoiding the pitfalls of relying solely on parameter counts.

---
Source: [FOSSHunter — Whichllm](https://www.fosshunter.com/tools/whichllm) — cite `synced_at` and the catalog revision when quoting metrics.
Interpretation rules: https://www.fosshunter.com/llms-full.txt