Google Veo 3.1 Lite
Google generative video model optimized for developers, available via Gemini API from ~$0.05/s.
Description
Google Veo 3.1 Lite is the budget variant of Google DeepMind's Veo 3.1 generative video model, launched on March 31, 2026 for developers who need high-volume, low-cost programmatic video generation. It's accessed via Gemini API and Google AI Studio under the identifier 'veo-3.1-lite-generate-preview' and offers text and image input, video output with audio, and cinematic control, optimized for the same speed as Veo 3.1 Fast at roughly half the price. It's currently in Preview status: it doesn't support 4K or the Extension feature of higher variants and caps text prompts at 1,024 tokens. The approximate price is $0.05 per second of generated video, with $300 of free credit for new Google Cloud accounts (~100 minutes of video). It's ideal when you want to integrate Veo into a product without paying the full model's cost.
Preview

Detailed Evaluation
Key advantages
Unbeatable per-second price at Google quality
Around $0.05 per second gives access to Veo 3.1 for roughly half the cost of Fast, ideal for prototyping and scaling.
Built-in native audio
Generates video with ambient sound and dialogue without having to orchestrate a separate audio pass.
Developer-focused API
Consumed from the Gemini API with the official SDK, integrates with Vertex AI, and fits into IaC pipelines.
Supports text and image as input
Allows image-to-video to keep visual consistency or re-animate stills, useful for branding.
High speed compared to other models
Matches Veo 3.1 Fast in generation time, enabling short iterations during prototyping.
Limitations to consider
Preview status, not GA
The model lives under the preview endpoint and can change, quotas and SLAs aren't final; risky for critical production.
No 4K or Extension
Videos cap at 1080p and can't be extended with the Extension feature that exists on full Veo 3.1.
Prompt limited to 1,024 tokens
Forces you to be precise; very detailed scenes need to be split or used with image-to-video.
Regional and quota availability
Being preview, there are region restrictions on Google Cloud and quotas that can cut generation at peak hours.
Standout Feature
Being the cheapest and fastest way in 2026 to programmatically access the Veo 3.1 family with native audio, at roughly $0.05/s and the same speed as Veo 3.1 Fast, explicitly designed for programmatic production use.
Comparison with Alternatives
Against Veo 3.1 Fast/Full it's cheaper but gives up 4K and Extension; against Kling or Wan 2.5 via Fal it's more expensive per second but brings Google's quality and audio; against Runway it has no visual editor, it's pure API to integrate into your product.
Ideal User
Developers and product teams integrating video generation inside their app, needing predictable per-second pricing and already working with Gemini API or Vertex AI. Also fits growth teams that generate programmatic video variants for performance marketing.
Learning Curve
With the Gemini SDK you can generate your first video in five minutes. The real difficulty lies in writing effective cinematic prompts within the token limit and managing quotas and retries.
Best For
- Products integrating video generation via API at low per-second cost
- Rapid prototyping and A/B testing of video prompts
- High-volume workflows like programmatic ads or creative variants
- Teams already established on Google Cloud and Vertex AI
- Cases where 1080p output with native audio is sufficient
Not Ideal For
- Projects that require 4K or the Extension feature for long clips
- Pipelines that can't depend on a Preview-stage model
- Users who prefer NLE-style creative interfaces over an API