GitHub Repository

Find the local LLM that actually runs and performs best on your hardware. Ranked by real, recency-aware benchmarks, not parameter count. One command, run it instantly.

2,608 starsPython

Find the best local LLM for your hardware, ranked by benchmarks

Name: Find the best local LLM for your hardware, ranked by benchmarks
Availability: InStock
Author: andyyyy64

by andyyyy64·May 15, 2026·283 points·68 comments

Visit Project View on HN

AI Analysis

●●SolidSolve My ProblemShip It

Ranks models by actual benchmark scores instead of just fitting the biggest model in VRAM.

Strengths

•Recency-aware scoring demotes stale leaderboards so new models beat older, larger ones.
•Simulates specific GPU configs like RTX 4090 to predict performance before buying hardware.
•Aggregates diverse benchmarks like LiveBench and Chatbot Arena into a single weighted score.

Weaknesses

•Relies on third-party benchmark data consistency which can vary wildly between sources.
•Doesn't actually run the models locally to verify speed, relying on theoretical estimates.