Google Veo 3.1
Google's video model for cinematic shots with matching sound; Google is moving video generation to Gemini Omni Flash.
Quality
Modality
video
Resolution
N/A
Access
closed
Fabian's Take
"When I need a short video clip for a presentation that genuinely looks like it came from a stock video site, I use Veo. The physics coherence is incredible."
Veo 3.1 is Google DeepMind’s video model, first released in October 2025 and extended with a cheaper Lite version in spring 2026. It makes clips of 4, 6 or 8 seconds in landscape or portrait, up to 4K, with sound that matches the picture.
Cinematic understanding
Veo’s strength is its understanding of cinematic language. You can prompt it with specific camera movements (like “aerial tracking shot” or “macro timelapse”) and it usually gets them right. It handles physics, like water or complex lighting, better than most competitors.
A model in transition
Google is moving video generation to Gemini Omni Flash, a newer model that also edits and extends clips. The Veo 3.1 preview in the Gemini API shut down on 22 October 2026, with Omni Flash named as the replacement; on Google’s Vertex AI platform the finished Veo 3.1 versions stay available until at least 17 November 2026. Veo stays the pick here until I’ve tried Omni Flash.
The Verdict
Best for: High-fidelity cinematic shots and realistic physics simulation.
Pros
- Strong physical coherence in complex scenes
- Understands advanced cinematography terms
- Generates sound to match the video, included in the price
- Up to 4K in the Standard version
Cons
- Developer access is narrowing: the Gemini API preview shut down on 22 October 2026 and Google points to Gemini Omni Flash; Vertex AI keeps Veo 3.1 for now
- Short clips: 4, 6 or 8 seconds
Specs
- Pricing Per second, sound included: Standard $0.40 (720p/1080p) or $0.60 (4K) · Fast $0.10–0.30 · Lite $0.05–0.08
- Cost Tier premium
- ⚡Speed Tier moderate
- License Proprietary
