MicroLLM Lab
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.