Google DeepMind open-sourced Gemma 4 on April 8 under Apache 2.0 license (full commercial use). Sizes: 2B/7B/34B, optimized for edge/on-device deployment (phone, IoT, auto). Performance near Llama 3, strong coding/reasoning/multimodal. Gemma 4 democratizes top-tier AI, expected to boom edge AI.
#Gemma4 #GoogleAI #OpenSourceAI #EdgeAI
