Announcing Gemma 3n preview: powerful, efficient, mobile-first AI- Google Developers Blog

Introduces Gemma 3n, a new open model built on a shared, mobile-first architecture optimized for on-device AI. Highlights innovations like Per-Layer Embeddings (PLE) for reduced RAM usage, enabling larger models on mobile devices. Details key capabilities including optimized on-device performance, many-in-1 flexibility with nested submodels, privacy-first offline readiness, expanded multimodal understanding (audio, image, video), and improved multilingual performance. Provides access through Google AI Studio and Google AI Edge for cloud-based and on-device development respectively.