Musubi Launches PolicyLM-1.7B for Real-Time Content Moderation

Musubi has unveiled PolicyLM-1.7B, a lightweight decision model designed for real-time content moderation, capable of applying content policies to messages in just 50 milliseconds. This model is positioned as a direct competitor to TypeSafe AI's Jev, which was released in September. Filip Jankovic, co-founder and chief AI officer of Musubi, emphasized the need for product teams to gain better insights into platform activity as content volume continues to grow exponentially. The GLiNER project, initiated in 2024, aims to enhance Named Entity Recognition, further highlighting the industry's focus on improving content moderation technologies. Competitors in this space include OpenAI and Amazon, both of which are also developing decision models for similar applications. The launch of PolicyLM-1.7B provides practitioners with a new tool for content moderation, enabling them to run the model independently and tailor it to their specific needs.