Musubi released PolicyLM-1.7B, a lightweight open-weight decision model for real-time content moderation, applying plain-English content policies to messages in under 50 milliseconds. Its cost and speed are close to mainstream social platform moderation classifiers, and it requires no retraining when policies change. Co-founder and Chief AI Officer Filip Jankovic said the model lets platform managers label content proactively in a scalable, customizable way.