Google DeepMind released Nano Banana 2 Lite and Gemini Omni Flash for developer integration. These lightweight models prioritize low-latency inference for edge devices. The update provides a smaller memory footprint without sacrificing core reasoning capabilities. Developers can now deploy these models on mobile hardware to reduce cloud dependency and operational costs.