AI article

I built an open-source LLM benchmarking tool with an AI judge — here's how it works

When I started running local LLMs in my homelab, I kept hitting the same problem: tokens per second...

Dev.to | Mar 13, 2026 | MegabytesNYC

Read the original article