Veo 3, Imagen 4 and Lyria 2 on Vertex AI: Google's enterprise generative media stack gets a coordinated refresh

At I/O 2026 Google brings three refreshed generative models to Vertex AI — for video, images and music. Imagen 4 is in public preview, Veo 3 stays in private preview with broader access coming in the next few weeks, and Lyria 2 is generally available. It is the first time the enterprise creative pipeline lands in sync across all three media.

Among Google's many I/O 2026 announcements there is one move clearly aimed at people building products on top of Google Cloud: the coordinated arrival of Veo 3 for video, Imagen 4 for images and Lyria 2 for music inside Vertex AI. This is not just a model bump — it is the first time all three formats land together in the same enterprise environment, with the same safety policies and the same release plan.

What is available and when

Imagen 4 is in public preview on Vertex AI: anyone with a Google Cloud project can start testing it now. Veo 3 stays in private preview with a wider opening announced for the coming weeks, a caution that fits the fact that video is where Google sees the highest abuse risk. Lyria 2 is generally available instead, which means SLAs, list pricing and enterprise support from day one.

Why it matters

The interesting part is not the individual models but the fact that an agency, a broadcaster or a brand can now plan a unified generative pipeline — storyboards with Imagen 4, animation with Veo 3, soundtrack with Lyria 2 — staying inside the Vertex AI perimeter, with the same credentials, the same billing and the same provenance controls (Google reiterates SynthID integration for generated content).

The second piece is the connection to Google Flow, the creative tool Google demoed in Mountain View and that is now landing on mobile. Flow taps the same models that run inside Vertex AI: prosumer and enterprise users work on the same technology base, a clear departure from the old pattern where the newest models stayed locked to consumer products for months.

What is still open

Version numbers remain a slippery topic: Google has already talked about Lyria 3 Pro in the April 2026 Gemini Drops for consumer music creation, while the Vertex AI release ships as Lyria 2. It is the usual split between product models and platform models, but the messaging stays confusing for anyone joining the ecosystem now. On Veo 3 the real issue is inference latency: long-video generation windows remain compute-expensive, and public pricing will only show up at GA.

← Back to all announcements