Meet your next
local AI model.

For conversations, code, and ideas worth exploring. Discover the models, check their memory needs, and find hardware that fits.

Memory fit is an estimate. Actual speed depends on your hardware and setup.

Catalog examples, not a quality ranking.

157curated models

27model families

53permissive licenses

0measured fit anchors

Latest record review: 2026-10-08. Dates vary by model.

Start with memory.

Estimated requirements at the default quantization. Counts follow your current filters.

Search by model or family, then use memory and task to reduce the field.

Run one measured benchmark and turn this catalog into a shortlist based on your actual memory and local inference speed.