Which build runs your model?

Model Quantization Context (tokens, optional) Target speed (tok/s, optional) Also show offload builds (GPU + system RAM, slower)

Choose a model to see compatible builds.

Speeds are estimates: measured means a curated public benchmark row matched the build, formula-derived means estimated from memory bandwidth. Real speed varies with engine, drivers and settings. Some links to retailers are affiliate links: we may earn a commission if you buy, at no extra cost to you. Affiliate disclosure · Data sources and estimates