- Google Flow is an AI filmmaking tool built on Veo, Imagen, and Gemini, designed around continuous scenes rather than one-off clips.
- Since Google retired Whisk on April 30, 2026, Flow has absorbed its Subject + Scene + Style image-remix features.
- Pricing ranges from a free tier with 50 daily credits to a $199.99/month Ultra plan with 25,000 monthly credits.
- Google has confirmed a next-generation video model is coming, but as of October 2026 it has not officially been branded “Veo 4” — Flow’s current flagship model remains Veo 3.1.
| Key Facts | |
|---|---|
| Developer | |
| Launched | May 20, 2025 |
| Core models | Veo 3.1 (video), Imagen (images), Gemini (creative assistant) |
| Free tier | 50 credits/day |
| AI Plus | $4.99/month — 200 credits/month |
| AI Pro | $19.99/month — 1,000 credits/month |
| AI Ultra | $99.99–$199.99/month — 10,000–25,000 credits/month |
| Absorbed tool | Whisk (shut down April 30, 2026) |
What Is Google Flow?
Flow is Google’s AI filmmaking platform, launched May 20, 2025 and built around Veo for video generation, Imagen for stills, and a Gemini-powered assistant for planning. Its core design goal is scene continuity: rather than generating isolated clips that don’t match from shot to shot, Flow is built to help users keep characters, objects, and settings consistent across a sequence of connected scenes — closer to how an actual film or ad gets edited together.
Its main tools include Scene Builder, which stitches individual clips into a continuous sequence; Ingredients to Video, which locks in a consistent character or object appearance across multiple generations; and camera controls for pan, tilt, and zoom that don’t require prompt-engineering workarounds to achieve.
How Flow’s Features Have Expanded
Google’s current flagship video model inside Flow is Veo 3.1, which added native audio generation, longer scene extensions, a Frame to Video tool that interpolates between a chosen first and last frame, and object add/remove editing. Following Google I/O 2026, Google rolled out Flow Agent — a Gemini-powered assistant inside Flow capable of multi-step planning and retaining project memory — to all Flow users globally.
Flow’s feature set also grew by absorption: in February 2026, Google Labs announced that Whisk, ImageFX, and Flow would merge into a single creative platform. Whisk’s signature capability — generating a new image by combining a Subject, a Scene, and a Style image rather than typing a text prompt — along with its Imagen-powered generation, is now built into Flow directly. Whisk itself stopped accepting new generations in early April 2026 and fully shut down on April 30, 2026, with any undownloaded images from its library permanently deleted.
Flow Pricing
Flow uses a credit system layered on top of Google’s existing AI subscription tiers. The free tier includes 50 credits per day. Google AI Plus, at $4.99/month, includes 200 credits per month. Google AI Pro, at $19.99/month, includes 1,000 credits per month. Google AI Ultra comes in two tiers — $99.99/month for 10,000 credits, or $199.99/month for 25,000 credits. Generation costs vary by quality setting: a “Fast” video costs roughly 20 credits (about 10 on the Ultra plan), while a higher-quality “Quality” generation costs roughly 100 credits.
Flow vs. Sora vs. Runway
| Tool | Entry Price | Top Consumer Tier | Model Access |
|---|---|---|---|
| Google Flow | Free (50 credits/day) | AI Ultra, $199.99/month | Google models only (Veo 3.1, Imagen) |
| Sora (OpenAI) | ChatGPT Plus, $20/month | ChatGPT Pro, $200/month | Sora 2 only |
| Runway | Standard, $15/month (625 credits) | Unlimited, $95/month | Multi-model — Runway’s own Gen-4.5 plus third-party models like Sora 2, Seedance, and Kling |
The practical trade-off: Flow is tightly bundled with Google’s own model stack and ecosystem (Gemini, Vertex AI), which keeps pricing competitive for 1080p output but limits users to Google’s models. Runway’s multi-model approach appeals to creators who want to switch between different video models — including Sora 2 itself — from one interface, and its annual billing can undercut Flow on 4K upscaling. Sora’s access is gated entirely behind a ChatGPT subscription rather than sold as a standalone product.
Is Veo 4 Available Yet?
As of October 2026, Google has not released a model officially branded “Veo 4,” despite ongoing speculation. Google’s own product pages still list Veo 3.1 as the company’s leading video generation model inside Flow, the Gemini API, the Gemini app, and Vertex AI. At Google I/O 2026, Google did announce a major new video-related model, but it was named Gemini Omni Flash rather than Veo 4 — a distinction worth noting for anyone searching specifically for a Veo 4 release date, since no such official release has been confirmed.
Who Should Use Flow
Flow is best suited to creators already inside Google’s ecosystem — someone making short-form ads, social clips, or storyboard previews who wants character and scene consistency across several generations without manually re-prompting for every shot. Scene Builder matters most for anyone stitching together a sequence longer than a single clip, since it keeps lighting, character appearance, and setting aligned from one scene to the next rather than treating each generation as an isolated image. Ingredients to Video solves a narrower but common problem: keeping a specific product, mascot, or character looking the same across a whole campaign, rather than having the model reinterpret it slightly differently every time. Teams that need to switch between several different video models — say, comparing Veo output against Sora 2 or Kling for a specific shot — are better served by Runway’s multi-model interface than by Flow, which is locked to Google’s own models.
Getting Started with Flow
New users can try Flow’s free tier — 50 credits per day — without a Google AI subscription, which is enough for a handful of short, lower-quality generations to evaluate Scene Builder and Ingredients to Video before committing to a paid tier. Anyone coming from the discontinued Whisk specifically for its Subject + Scene + Style image remixing should look for that same workflow inside Flow’s image tools rather than searching for a separate app.
Frequently Asked Questions
Is Whisk still part of Flow? Yes — Whisk’s Subject + Scene + Style remixing is now built into Flow after Whisk itself shut down on April 30, 2026.
Can I use Flow without paying? Yes, Flow’s free tier provides 50 credits per day, though higher-quality generations consume credits faster.
Does Flow support models other than Google’s own? No — unlike Runway, Flow only runs Google’s Veo and Imagen models.
What is Flow Agent? It’s a Gemini-powered creative assistant inside Flow, rolled out globally after Google I/O 2026, that can plan multi-step projects and retain memory of a project’s context across a session.
Bottom Line
Flow has become Google’s single consolidated hub for AI-generated video and imagery, absorbing Whisk’s remix tools and continuing to build out Veo 3.1’s scene-continuity features through additions like Flow Agent. The rising search interest likely reflects both organic growth in Flow’s user base and the lingering confusion from Whisk’s April shutdown sending its former users toward Flow. Anyone expecting a separately branded “Veo 4” should note that, as of this writing, Veo 3.1 remains Google’s current production model.
