Why 8/10: Top pick for AI video: Veo 3.1 generates short, high-fidelity clips with native sound and dialogue, and it is reachable through consumer plans and a pay-per-second API. Points off for short clip lengths, several overlapping credit systems, and Google now pushing a newer Gemini Omni model alongside it.
By ToolSift editorial. Last checked . Prices change often; "check site" means verify with the vendor.
What it does
Veo is Google DeepMind's video generation model. DeepMind's Veo page describes Veo 3.1 as "our leading video generation model, designed to empower filmmakers and storytellers", and highlights native audio: sound effects, ambient noise and dialogue are generated with the picture rather than added afterwards. Google's developer docs say Veo 3.1 makes 8-second videos at 720p, 1080p or 4K with audio, supports portrait (9:16) and landscape (16:9), can extend earlier Veo videos, can generate from a first and last frame, and accepts up to three reference images to steer the content.
There are three ways in. The Gemini app gives video generation to paying Google AI subscribers. Google Flow is Google's AI film studio: Flow's help pages list Veo 3.1 Lite, Fast and Quality, with 4, 6 and 8-second clips for most modes and 8-second clips for extensions and reference-image generation. Developers can call Veo 3.1 through the Gemini API, paying per second of output.
Flow now also offers Gemini Omni Flash 1.1, a newer Google model that can generate 10-second clips and edit uploaded or generated video, which Veo 3.1 cannot do in Flow. Veo remains Google's headline realism model, but if you need editing of existing footage, compare the two inside Flow.
Against our runners-up: Kling AI is a strong choice for longer clips and lower prices per generation, and Runway offers a broader editing toolkit for video professionals. Veo's edge is realism with sound: dialogue, ambient noise and effects arrive with the picture, which saves a separate audio step for short hero shots. If your project is mostly editing existing footage, compare Gemini Omni Flash in Flow, or Runway, before committing credits.
Our verdict
Veo 3.1 is still the safe default for short, realistic AI video with sound. Google DeepMind presents it as its leading video model, and it generates 4, 6 or 8-second clips with natively generated audio, including dialogue, in landscape or portrait. You can use it in the Gemini app, in Google's Flow studio with monthly credits, or pay per second through the Gemini API. Expect to build longer pieces shot by shot and to learn a few credit systems. It is the pick for a hero shot or a talking subject, not for one-click long-form video.
✅ Pick it if You need a short, high-fidelity, realistic shot with sound or dialogue, or you want pay-per-second API access for a product.
⛔ Skip it if You need long multi-shot sequences in one generation, video-to-video editing with Veo itself, or one simple flat price.
Pricing: As of 19 Oct 2026: Gemini API (paid tier only, per second of video with audio): Veo 3.1 Standard $0.40 (720p/1080p) or $0.60 (4K); Fast $0.10 (720p), $0.12 (1080p), $0.30 (4K); Lite $0.05 (720p), $0.08 (1080p). In Flow, everyone gets 50 credits a day; Veo 3.1 Lite costs 10 credits, Fast 20 and Quality 100 per generation for non-Ultra users. Google AI plans add monthly Flow credits; UK prices shown were AI Plus £4.49/month (200 credits), AI Pro £18.99/month (1,000 credits) and AI Ultra from £79.99/month (10,000 or 25,000 credits). Check site.
Veo is built by Google DeepMind and sold through Google's consumer AI plans, Google Flow and the Gemini API (Google LLC).
What to watch
Clips are short. Veo 3.1 generates up to 8 seconds per clip; longer videos mean extending or stitching clips, and Flow says Veo 3.1 extensions must use the Lite model.
Three credit systems: Gemini app limits, Flow credits and Gemini API billing are separate. Flow's help page says non-subscribers may not be able to generate video at peak times (roughly 2–5 PM UTC).
Free-tier API use is not available for Veo, and Google's pricing page notes that free-tier data is used to improve its products, while paid-tier data is not.
Feature support differs by variant: in Flow, Veo 3.1 Quality does not support reference images or extensions, and none of the Veo 3.1 variants support video-to-video editing.
Google is also promoting Gemini Omni Flash for video in Flow (10-second clips and editing). Check which model a given feature uses before you budget credits.