Google Flow pricing (2026): plans, credits & video costs

Flow is Google's AI creative studio for images and video. It combines Veo 3.1, Nano Banana, Gemini Omni and project tools in one workspace. Prices, eligibility, taxes and model access vary by country and account, so confirm the live plan and credit estimate before generating.

Magic Hour publishes this guide and offers separate image and video generation tools. Flow product, model, plan, and credit claims are based on the linked official Google sources; this is a documentation review rather than a controlled comparison with Magic Hour.

What is Google Flow?

Flow is an application, not a single video model. Veo generates video; Nano Banana handles image generation and editing; Gemini Omni supports multimodal video creation and editing. Flow adds project assets, scene assembly and tools around those capabilities.

Google introduced Flow on May 20, 2025. Later updates expanded the creative studio. The original launch should not be confused with a 2026 feature release. Google’s May 2026 announcement also describes Flow Music, a related music creation tool.

Google Flow pricing at a glance

Google's Flow plan display and credit guide were checked September 13, 2026. These are US-dollar list prices and allowances; regional prices, eligibility and taxes can differ.

How many Flow credits does one video cost?

Google charges per generation, not per request. One request can create more than one video, so inspect the generation count before approving it. The current credit guide lists:

• Veo 3.1 Lite: 10 credits per 4-, 6-, or 8-second generation for non-Ultra subscribers and 5 for Ultra subscribers.
• Veo 3.1 Fast: 20 credits per 4-, 6-, or 8-second generation for non-Ultra subscribers and 10 for Ultra subscribers.
• Veo 3.1 Quality: 100 credits per 8-second generation for every plan.
• Gemini Omni Flash at 720p: 7 credits for 4 seconds, 10 for 6, 12 for 8, and 15 for 10 seconds.
• Gemini Omni Flash video editing: 40 credits per generation; 1080p upscaling is included for subscribers, while 4K upscaling costs Ultra users 50 credits.

At the listed rates, 50 daily credits can cover at most five non-Ultra Veo 3.1 Lite generations, two Fast generations with 10 credits left, or no Quality generation until enough credits are available. These are allowance calculations, not promises of five usable clips; retries, multiple outputs, edits and upscaling consume more credits.

Unused daily credits do not roll over, and monthly subscription credits refresh at the start of the billing cycle. Confirm the estimate shown before generating. Model, duration, resolution, output count, extensions, editing and upscaling all affect the real cost per usable video.

If free export matters more than Flow's model set, compare the current free AI video generators without watermarks.

Flow subscription credits are different from API billing

A Flow subscription pays for usage inside the consumer application. It is not a general-purpose Gemini or Vertex AI API credit balance. Developers should compare the exact model identifier, resolution and audio setting on the Gemini API pricing page. Do not apply a Flow credit conversion to an API quote.

How to make your first video in Flow

1. Open Google Flow and sign in with an eligible Google Account.
2. Start a project and choose the input mode available to your account: text, frames or other references.
3. Describe one shot, including the subject, action, setting and camera movement.
4. Check the selected model and credit estimate, then generate a short draft.
5. Review movement, faces, text and audio before extending the shot or assembling a sequence.

Example prompt to adapt: “A red ceramic mug on a wooden desk beside a window. Steam rises slowly. The camera makes a gentle push-in. Soft morning light, no text.” This is a suggested prompt, not a benchmark result.

Using Veo specifically? Follow the Veo 3.1 first-clip workflow for model selection, an example prompt and a concrete credit budget.

Which Flow workflow should you use?

A model’s availability does not mean every feature works with it. Use the modes shown for the selected model instead of assuming that all references, aspect ratios and durations are interchangeable.

When is an alternative useful?

Choose Flow when you want its available Google models and project workspace. For a direct browser workflow, compare Magic Hour Image-to-Video or Text-to-Video. For existing footage, use the specific operation you need, such as face swap or lip sync. Our image-to-video guide compares the generation workflows and their limits.

Compare a permitted sample using the same brief and intended output format. Include retries, export restrictions and commercial terms in the decision. A larger credit balance alone does not establish lower cost or better output.

Google Flow or Magic Hour: the fastest decision

Choose Google Flow when Veo access, Scenebuilder, or Google’s filmmaking workspace is the main requirement. Choose Magic Hour when you want one browser workspace for generation plus follow-on tasks such as face swap, lip sync, talking photos, and other image, video, and audio tools.

The fair comparison is one finished shot, not one generation. Write a brief with subject, action, setting, camera movement, duration, orientation, and audio requirement. Generate the same shot, then record candidates used, credits or billed seconds, wait time, export resolution, watermark, and manual correction time.

Reject outputs that miss a required product detail, change the subject, break continuity, or create unusable audio. Divide total generation cost by accepted shots. This prevents a low per-run price from looking cheap when the workflow needs many retries.

If your starting point is a real image, test Magic Hour Image-to-Video. If the shot begins as text, test Text-to-Video. Keep Flow subscription credits separate from Gemini API billing.