Google Veo 3 is 📝Google's third-generation video generation model, announced at Google I/O on May 20, 2025, and the first widely available model to generate video with synchronized native audio — dialogue, sound effects, and ambient sound — from a text prompt.
Veo 3 was developed by 📝DeepMind and launched in the 📝Gemini app and Google's Flow filmmaking tool, where its realistic, fully voiced clips quickly spread across social media and prompted debate about how easily AI video could pass for real footage. It reached the Gemini API in 2025 and was succeeded by Veo 3.1 on October 15, 2025, which added richer audio and more creative controls. Google shut down the Veo 3 API models on June 30, 2026, directing developers to Veo 3.1.
Key Facts
- Developer: Google DeepMind
- Announced: May 20, 2025, at Google I/O
- Notable for: native audio generation alongside video
- API models:
veo-3.0-generate-001andveo-3.0-fast-generate-001, shut down June 30, 2026 - Predecessors: Veo (May 2024) and Veo 2 (December 2024)
- Successor: Veo 3.1 (October 15, 2025)
FAQ
What made Veo 3 different from earlier video models?
Veo 3 generated synchronized audio, including speech and sound effects, together with the video, where earlier models produced silent clips.
Where do Veo models fit in Google's lineup?
Veo is Google's dedicated video model line; every Veo, Gemini, and Nano Banana release is indexed in the 📝Gemini Model Glossary.
Is Veo 3 still available?
Not in the Gemini API, which shut down the Veo 3 models on June 30, 2026. Google's current video models are Veo 3.1 and Gemini Omni Flash.