JevMade hello@JevMade.com
← Back to experiments

Benchmarks & research

typesafe-local

typesafe-local asks a local MLX model typed questions and returns calibrated probabilities without generating or parsing prose.

Source screenshot of typesafe-local
SOURCE SCREENSHOT · source ↗ · captured 2026-09-21Full screenshot ↗

What it does

Built for Apple Silicon, it reports roughly 200 milliseconds for four questions over one document on an M4 Pro.

Maker-reported (not independently measured by JevMade): Four questions over one document: ~200 ms on an M4 Pro · Qwen3-1.7B, full validation split, bootstrap 95% intervals:

Primitives
choice, score, noul
Platform
Python
Added
Project created
GitHub stars
5 · snapshot 2026-09-18T22:58:45Z

Source checked 2026-09-19 — opened the GitHub repository directly.