Google I/O 2025: Major AI Announcements Unveiled

Google’s 2025 I/O event focused heavily on AI, introducing new tools and updates across its Gemini, Imagen, and Veo models, as well as launching a new AI filmmaking platform.

Gemini 2.5 (Pro & Flash) Supports 24+ languages with text-to-speech and expressive voices Improved reasoning, coding, multimodal understanding, and 1 million token context window. Pro model adds "Deep Think" mode for complex reasoning and coding (in testing) Flash model is available now in preview via the Gemini app Pro model commercial release coming soon Enhanced security against indirect prompt injections.

Imagen 4 Generates images in up to 2K resolution Better text accuracy in posters, cards, and comics Available now across Gemini, Google Workspace, Whisk, and Vertex AI.

Veo 3 Advanced AI video generation model Creates video with sound, dialogue, and background noise Available now for Google AI Ultra subscribers (US) and Vertex AI enterprise users Veo 2 gets updates too: Camera movement, object editing, style control, and outpainting.

Flow – AI Filmmaking Tool New tool that combines Veo, Imagen, and Lyria Creates cinematic scenes from text prompts Offers control over camera motion, angles, perspectives Allows editing and extending of generated clips Available now for Google AI Pro and Ultra subscribers (US).