Google debuts new AI models, personal AI agents in effort to keep pace with OpenAI and Anthropic

Google announced new AI models at its I/O conference, including Gemini 3.5 Flash (a faster, cheaper version of its main model) and Gemini Spark (an AI agent that can take actions in connected apps on a user's behalf). The company also introduced Omni, a world model (AI trained to simulate and predict physical environments) that can edit videos and generate realistic imagery, as Google competes with rivals like OpenAI and Anthropic.