Google launches Veo 3.1 Lite in Gemini API at $0.05 per second

• Google has shipped veo-3.1-lite-generate-preview in the Gemini API and AI Studio, with pricing that starts at $0.05 per second at 720p and $0.08 per second at 1080p launch thread model pricing.
• The biggest change is economic, not architectural: Veo 3.1 Lite lands at less than half the cost of Veo 3.1 Fast, while Google says it runs at the same speed Google announcement pricing table.
• Lite supports text-to-video and image-to-video, landscape and portrait output, and short 4, 6, or 8 second clips, which makes it a better fit for high-volume ad, social, and product-video workloads than for premium one-off renders feature list developer notes.
• Google is also cutting Veo 3.1 Fast pricing on April 7, which means teams already testing Fast should rerun their cost models instead of treating this as a Lite-only story April 7 price cut pricing recap.

The useful reveals are straightforward. Google’s launch post frames Lite as the high-volume tier, the model page confirms the API SKU and the missing 4K output, the video docs show that first-and-last-frame generation is part of the wider Veo 3.1 feature set, and the release notes make clear this is now an official Gemini API launch, not an AI Studio-only experiment.

Google did not sneak out a smaller Veo variant, it reset the entry price for API video.

The official price table puts Veo 3.1 Lite at $0.05 per second for 720p and $0.08 per second for 1080p. The same chart shows Veo 3.1 Fast at $0.15 today, dropping to $0.10 for 720p and $0.12 for 1080p on April 7, while the full-quality tier remains far higher at $0.40 and up pricing table. The Gemini pricing page backs that up.

For engineering teams, that changes what is reasonable to automate. At $0.20, a 4 second 720p clip moves into the range where you can test multi-variant creative generation, generate rough cuts for approval flows, or attach motion previews to catalog content without treating every API call like a premium asset render.

The launch is conservative in the right way. Google kept the familiar Veo 3.1 surface area and stripped back the expensive parts.

According to the launch post, Lite supports text-to-video and image-to-video, audio in output, landscape 16:9 and portrait 9:16, plus 4, 6, and 8 second durations. The model docs add the hard limit that matters most: no 4K output, and no Extension support.

That combination tells you what Google thinks this tier is for. Lite is aimed at rapid iteration and volume workloads where 720p or 1080p is enough, not at long-form cinematic generation.

One of the more interesting details is how fast the surrounding tooling filled in around Lite.

Google’s video generation docs describe Veo 3.1 controls for aspect ratio, direct resolution selection, image-driven generation, and first-and-last-frame video creation. Then fal exposed Lite almost immediately with separate endpoints for text-to-video, image-to-video, and first-last-frame-to-video fal endpoints.

That matters because the real adoption bottleneck for video APIs is often orchestration, not model access. Once hosted platforms package the same model into narrower endpoints, teams can slot it into ad pipelines, product demo generators, or template-based asset systems with less glue code.

The most concrete outside-the-launch-day signal came from an early tester who said Lite works especially well for "stock video" style footage, where vibe and scene structure matter more than exact on-screen text stock footage claim.

That feels right for this price tier. If you need lots of short atmospheric clips, B-roll, social inserts, or campaign variants, Lite is probably the new default starting point. If you need pristine 4K output or are already tuned around Fast or Quality, the April 7 price cut means the decision is less about whether to downgrade and more about where each tier belongs in your pipeline.

Google’s Gemini API changelog lists March 31 as the launch date for Veo 3.1 Lite Preview and explicitly calls it the company’s most cost-efficient video generation model for rapid iteration and high-volume applications.

That is the real signal to bookmark. Google now has a clearer three-step pricing ladder for generated video, and Lite is the tier most likely to turn video generation from an occasional demo feature into something product teams can afford to run continuously.