Running AI models on real, local GPUs — one honest test at a time.
Watch on YouTubeEvery test follows the same strict rules so you can trust the numbers.
Every model is tested with the exact same prompt so results are directly comparable.
All benchmarks run on a single, fixed hardware setup — no variable equipment.
Only one variable is changed per test — everything else stays constant.
The model is reloaded before every single run to eliminate caching bias.
Every prompt and result file is published in the open repository for full transparency.
Every benchmark runs on the same machine so the only thing changing is the model.
Fixed benchmarking setup
PC Magik Lab is a YouTube series dedicated to running AI models on real, local GPUs — with full transparency and honest benchmarks.
Each episode puts models through the same rigorous testing process: identical prompts, identical hardware, and a full model reload before every run.
The goal is simple — give you trustworthy, comparable data about how AI models actually perform on consumer-grade hardware.
All prompts, configurations, and result files are published openly so anyone can verify the results.