Google's Veo 3 Generates Video and Audio Together — and That Could Change Everything

• Google's video generation model just got a major upgrade
• Announced at Google I/O, Veo 3 can combine audio and video in a single output
• For now it's an Ultra-tier, US-only feature

AI video tools like Sora and Pika can already turn out unsettlingly realistic clips, and with enough effort you can chain those clips into a short film. The one thing they can't do is generate audio at the same time. Google's new Veo 3 model can — and that could be a game changer.

Unveiled on Tuesday at Google I/O 2025, Veo 3 is the third generation of Google's powerful Gemini video model. Give it the right prompt and it produces video complete with sound effects, background noise, and dialogue. Google offered a brief demonstration: a CGI-grade animation of animals talking in a forest, with sound and picture in perfect sync. If that demo translates to real-world use, it represents a remarkable tipping point for AI content generation.

"We're emerging from the silent era of video generation," said Google DeepMind CEO Demis Hassabis on a press call.

Lights, camera, audio

He isn't wrong. Until now, no other AI video model could deliver synchronized audio — or any audio at all — alongside its video. It's still unclear whether Veo 3, which like its predecessor Veo 2 should be capable of 4K output, beats current quality leader OpenAI Sora on raw visual fidelity; Google has previously claimed Veo 2 is good at producing realistic, consistent movement. Either way, the ability to output what look like fully produced clips — video and audio — may make Veo a far more attractive platform overnight.

It goes beyond dialogue. In film and TV, background noise and sound effects are usually the work of Foley artists. Now picture describing to Veo the sounds you want layered behind the action and getting back the whole package — video, dialogue, and effects. That's work that takes animators weeks or months. In its release on the model, Google suggests you tell the AI "a short story in your prompt, and the model gives you back a clip that brings it to life." If Veo 3 can follow prompts and output minutes — eventually hours — of consistent video and audio, the first animated feature generated entirely through Veo may not be far off.

Veo 3 is live today, available in the US as part of the new Ultra tier ($249.99 a month) in the Gemini app and through the new Flow tool.

Google also rolled out a few updates to Veo 2, including generating video from reference objects you provide, camera controls, outpainting to convert portrait to landscape, and object add and erase.