Veo 3.1

Google

Google DeepMind's Veo 3.1 video-generation model, released in paid preview on October 15, 2025 across the Gemini API, Google AI Studio and Vertex AI. It generates video with native audio from a text prompt, a starting image, or a first and last frame. Google lists the Gemini API preview codes (veo-3.1-generate-preview, veo-3.1-fast-generate-preview and veo-3.1-lite-generate-preview) as shutting down on 22 October 2026 and names gemini-omni-1.1-flash as the replacement.

Full Specifications

All specsBest valueCheapestHeavy workloadMax capabilityDifferences only

	Veo 3.1Google $0.40/sec (720p/1080p) or $0.60/sec (4K); Fast tier from $0.10/sec
Identity
Developer	Google
Released	Oct 2025
Status	Deprecated - shutting down 22 October 2026 on the Gemini API (veo-3.1-generate-preview, veo-3.1-fast-generate-preview, veo-3.1-lite-generate-preview); replacement: gemini-omni-1.1-flash. Vertex AI's GA veo-3.1-generate-001 lists retirement November 17, 2026 or later.
Licence	Proprietary
Self-hostable	No
Cost
Price per second	Standard tier: $0.40/sec (720p or 1080p), $0.60/sec (4K); Fast tier: $0.10/sec (720p), $0.12/sec (1080p), $0.30/sec (4K)
Cheapest plan	No subscription plan for direct Gemini API access - pure pay-as-you-go per second (see Price per second); on the consumer side, the cheapest paid route into video generation is Google AI Plus ($4.99/mo, 200 Google Flow credits/mo) for the Gemini app's built-in video generation, though Google's separate Flow video tool includes some free daily credits without any paid plan - see Free tier
Free tier	Not offered via the Gemini API ("Not available" on the pricing page); the Gemini app's own chat-integrated video generation requires at least Google AI Plus; Google's separate Flow creative tool (a dedicated video/image app) does include Veo 3.1 access on a free "Without a Google AI Subscription" tier (50 daily Flow credits)
Capacity
Max duration	8 seconds per single generation (selectable durations of 4, 6, or 8s); the separate Extend feature adds 7 seconds per call, repeatable up to 20 times, for a maximum stitched length of roughly 148 seconds ("a minute or more" per Google's own launch announcement)
Max resolution	720p (default) or 1080p at any of the supported durations; 4K only at the 8-second duration; all three resolutions are priced separately on the Standard and Fast tiers (see Price per second) - the separate Veo 3.1 Lite tier does not support 4K output at all
FPS	24fps
Capability
Text-to-video	Yes - primary interface is a natural-language text prompt (prompt in, video+audio out)
Image-to-video	Yes - a supplied image can be used to start the video (first-frame conditioning); also supports "first and last frame" interpolation (generate the transition between two supplied frames, with accompanying audio) and up to 3 reference images of a single person, character, or product for image-based direction
Native audio	Yes - natively generates synchronized audio (dialogue, sound effects, ambient noise) alongside video in a single call; per Google's own launch announcement, Veo 3.1 generates "richer native audio, from natural conversations to synchronized sound effects" than Veo 3
Camera controls	Prompt-based only - no dedicated camera-control API parameter; camera movement (POV shot, aerial view, tracking drone view, etc.) is specified via natural-language prompt text
Rights
Commercial rights	Yes - Google's Gemini API Additional Terms of Service state "Google won't claim ownership over that content," covering generated outputs generally (not video-specific wording), subject to the Generative AI Prohibited Use Policy
Access
API	Yes - Gemini API (generateVideos / predictLongRunning), Google AI Studio, and Vertex AI
Chat app	Yes - per Google's own launch announcement, "available in paid preview via the Gemini API, Google AI Studio, Vertex AI, the Gemini app, and Flow" (Google's dedicated AI video/creative tool)
Cloud marketplaces	Not offered - only available via Google's own Gemini API/Vertex AI; not found on AWS Bedrock or Microsoft Foundry (Azure)
Fine-tuning	Not offered - Google's own fine-tuning guide states no current Gemini API model supports tuning ("we no longer have a model available which supports fine-tuning in the Gemini API or AI Studio") and does not mention Veo at all