Google DeepMind launched Nano Banana 2 Lite and Gemini Omni Flash for developers. These models prioritize low-latency inference and efficiency for edge computing. The update enables faster local processing on mobile devices. Practitioners can now deploy high-speed multimodal capabilities without relying on heavy cloud infrastructure for every single request.