Meta's Muse Spark Open license counts distilled models as derivatives

aidenclarke_12 · reddit · 2026-11-04

A Reddit post urges developers to read the Muse Spark Open license before integrating it. The model is 816B (Beam: 501B), so for most people with a 24GB card, NVIDIA's MIT-licensed bolt was the only locally runnable big drop that week.
Big open-weight models mostly get used via APIs anyway, but the license doesn't care which route you take: Muse Spark Open requires showing "built with muse spark open" in your product, docs, or settings menu, and any company with over $20M revenue in a 12-month window must sign a separate deal with Meta before offering it as MaaS.
The easy-to-miss part: the license counts models trained or distilled on Muse outputs as derivatives, so a small model fine-tuned on its answers carries the same attribution rule. Internal use and non-MaaS app integrations stay outside the revenue clause.

More from Models