PC Magik Lab

Local models, real GPUs.

Watch on YouTube Results Repo

How We Measure

Same prompt across all tests

Same hardware rig

One variable changed per test

Model reloaded before every run

Every prompt and result file published

The Hardware

Testing Rig

NVIDIA RTX 3090 24 GB

Model Server

LM Studio

About the Series

PC Magik Lab is dedicated to exploring the practical side of local Large Language Models. We skip the marketing hype and focus on actual performance on consumer-grade hardware, providing transparent data for enthusiasts and developers who want to run AI on their own terms.