MicroLLM Lab – Try 7 tiny LLM's in the browser
MicroLLM Lab runs seven tiny LLMs (down to 135M parameters) in the browser with JavaScript-written objective benchmarks measuring speed and pass rates.
MicroLLM Lab is a browser-based tool for running and comparing seven small language models, including models as small as 135M parameters. Users write benchmarks in JavaScript that are evaluated in the page origin, with objective checks (regex and exact tokens) rather than writing quality. The tool measures tokens/s speed and accuracy pass rates, and all measurement data stays on the local machine.
- Runs seven tiny LLMs (down to 135M parameters) in-browser
- Benchmarks written in JavaScript, evaluated in local origin
- Measures sustained tokens/s and objective pass rates
- All run data stays on the local machine
From the original article
1 min read at stateofutopia.comObjective checks (regex / exact tokens), not writing quality. A 135M model is allowed to fail — that is the measurement. Pick models, then run.
By logicallee · stateofutopia.com
A short excerpt; the article and its images belong to the publisher. Summary, key facts and assessment above are AI-written from the article and can be wrong.