Google's fastest omni-capable AI model — ultra-low latency multimodal intelligence across text, image, audio, and video inputs for real-time applications, live interactions, and high-throughput developer pipelines