Vibeleaderboard
Index / app

MicroLLM Lab

stateofutopia.com
Visit stateofutopia.com
Category
AI Tools
Rank
No. 2709Tools index
Pricing
Free
Type
APP
Date

About

MicroLLM Lab runs 4-bit quantized small language models (roughly 25M to 360M parameters) in the browser using WebGPU. Users can chat with a loaded model, run speed and accuracy benchmarks, compare models, and generate a shareable performance certificate. Models are cached in the browser's IndexedDB, so prompts stay on the device.

Tags

webgpusmall-language-modelson-devicebenchmarkquantizationbrowser

Comments (0)

No comments yet

Editorially curated, with community endorsements as a secondary signal. Corrections welcome.