Google AI Studio Veo Shutdown: Where to Use the Veo 3.1 API Now
Veo 3.1 preview left Google AI Studio and the Gemini API on 2026-10-22. Where Veo 3.1 still runs, Gemini Omni Flash as the replacement, and prices.
Last Updated:
本文暂无中文版，以下为英文原文。
Short answer: Google shut down the Veo 3.1 preview models (veo-3.1-generate-preview, veo-3.1-fast-generate-preview and veo-3.1-lite-generate-preview) in the Gemini API on 2026-10-22. With a Google AI Studio key, use Gemini Omni Flash (gemini-omni-1.2-flash) instead: 3 to 15 second videos with native audio for about $0.07 per second at 720p. To keep using Veo 3.1, call the GA models (veo-3.1-generate-001 and veo-3.1-fast-generate-001) on Google Cloud's Gemini Enterprise Agent Platform, where Veo 3.1 with audio costs $0.40 per second at 720p and 1080p.
Key takeaways
- AI Studio no longer serves Veo. Google's Veo page in the Gemini API docs now starts with the shutdown notice and points to Omni Flash or Google Cloud.
- Omni Flash is cheaper and longer. About $0.07 per second at 720p against $0.40 for Veo 3.1, and up to 15 seconds per clip against 8.
- Veo 3.1 GA runs only in
us-central1on Google Cloud, with 4, 6 or 8 second clips. gemini-omni-1.1-flashroutes to 1.2 on 2026-12-03, billed at the 1.2 rates.
What changed
The Veo page in the Gemini API docs now says: "Veo 3.1 Preview (veo-3.1-generate-preview, veo-3.1-fast-generate-preview, and veo-3.1-lite-generate-preview) in the Gemini API was shut down on October 22, 2026." It asks developers to migrate to Gemini Omni Flash or to the GA Veo models on the Gemini Enterprise Agent Platform. Veo 3 and Veo 3 Fast are marked deprecated on the same page.
So a search for "Google AI Studio Veo" or "Veo 3 API" today leads to two paths:
| Gemini Omni Flash | Veo 3.1 | Veo 3.1 Fast | |
|---|---|---|---|
| Where | Gemini API, Google AI Studio | Gemini Enterprise Agent Platform | Gemini Enterprise Agent Platform |
| Model code | gemini-omni-1.2-flash | veo-3.1-generate-001 | veo-3.1-fast-generate-001 |
| Clip length | 3 to 15 seconds; extension to 30 | 4, 6 or 8 seconds | 4, 6 or 8 seconds |
| Resolution | 360p, 720p, 1080p and 4K (upscaled) | 720p, 1080p, 4K | 720p, 1080p |
| Aspect ratio | 16:9, 9:16 | 16:9, 9:16 | 16:9, 9:16 |
| Audio | Native audio | Generated audio | Generated audio |
| Inputs | Text, images (up to 6 references), reference videos, a video to edit | Text, image, first and last frame, reference images | Text, image, first and last frame, reference images |
| Editing | Multi-turn editing in natural language | Video extension | Video extension |
| Region | Gemini API | us-central1 | us-central1 |
Sources: the Gemini Omni Flash model page and the Veo 3.1 model page, read on 2026-11-15. Veo 3.1 Lite (veo-3.1-lite-generate-001) is in preview on Google Cloud, without reference images.
Prices per second of video
Gemini Omni Flash bills output tokens at $12.00 per million on the Gemini API paid tier, and it is not on the free tier. Google converts that to an effective price per second (Gemini API pricing):
| Resolution | Tokens per second | Price per second | 10-second clip |
|---|---|---|---|
| 360p | 1,931 | $0.023 | $0.23 |
| 720p | 5,792 | about $0.07 | about $0.70 |
| 1080p (upscaled) | 8,688 | $0.104 | $1.04 |
| 4K (upscaled) | 17,376 | $0.209 | $2.09 |
Input costs $1.50 per million tokens for text, images and video.
Veo on the Gemini Enterprise Agent Platform, per second of video (Google Cloud pricing):
| Model | Video with audio | Video only |
|---|---|---|
| Veo 3.1, 720p and 1080p | $0.40 | $0.20 |
| Veo 3.1, 4K | $0.60 | $0.40 |
| Veo 3.1 Fast, 720p | $0.10 | $0.08 |
| Veo 3.1 Fast, 1080p | $0.12 | $0.10 |
| Veo 3.1 Fast, 4K | $0.30 | $0.25 |
| Veo 3.1 Lite, 720p | $0.05 | $0.03 |
| Veo 3.1 Lite, 1080p | $0.08 | $0.05 |
An 8-second Veo 3.1 clip with audio at 1080p costs $3.20. The same clip from Veo 3.1 Fast costs $0.96, and from Omni Flash about $0.83.
Generate a video with Gemini Omni Flash
Create a key in Google AI Studio, set it as GEMINI_API_KEY, and install the Google Gen AI SDK (pip install google-genai). Omni Flash uses the Interactions API (Omni Flash guide):
import base64
from google import genai
client = genai.Client() # reads GEMINI_API_KEY
interaction = client.interactions.create(
model="gemini-omni-1.2-flash",
input="A drone shot of a mountain landscape at sunrise.",
response_format={
"type": "video",
"resolution": "1080p",
"aspect_ratio": "16:9",
"duration": "8s",
},
)
with open("mountain.mp4", "wb") as f:
f.write(base64.b64decode(interaction.output_video.data))duration takes whole seconds from "3s" to "15s", and the default is "10s". For files larger than 4 MB, set delivery to "uri" in response_format and download the video from the returned link. To edit the result, send a new interaction with previous_interaction_id and describe the change.
Limits to plan for
- No audio input. Audio files and audio references are not supported in the current API.
- People. Uploading images of certain recognizable people is not supported. In the EEA, Switzerland and the UK, images of minors and editing of uploaded videos are not supported.
- Extension. Each turn adds 3 to 10 seconds to the end of a video, up to 30 seconds in total.
- Watermark. Every video carries an invisible SynthID watermark.
- Prompts. English is fully supported. System instructions, temperature and negative prompts are not.
Migrating from Veo 3.1 preview
- Choose the target. Omni Flash if you want the Gemini API key, longer clips and editing. Veo 3.1 GA on Google Cloud if you need Veo's output or already run on Google Cloud.
- Change the call. Veo used a long-running
generate_videosoperation. Omni Flash usesclient.interactions.createwithresponse_format. Veo on Google Cloud needs a Google Cloud project, theus-central1location and the GA model codes. - Check durations. Veo 3.1 GA makes 4, 6 or 8 second clips. Omni Flash makes 3 to 15 second clips, and first-and-last-frame interpolation supports 3 to 10 seconds.
- Re-run your prompts. The models differ, so compare outputs on your own prompts before you switch production traffic.
Other video models on one key
AIsa does not serve Veo or Gemini Omni Flash. If you want video generation on the same key as your language models and data APIs, AIsa lists these video models in the model catalog, priced on 2026-11-15:
| Model on AIsa | 720p per second | 1080p per second |
|---|---|---|
Wan 3.0 Video (wan3.0-video) | $0.037 | $0.074 |
Wan 2.7 text-to-video (wan2.7-t2v) | $0.060 | $0.100 |
HappyHorse 1.1 text-to-video (happyhorse-1.1-t2v) | $0.098 | $0.126 |
Seedance 2.5 (dreamina-seedance-2-5) | Token-based, about $0.23 | Token-based, about $0.57 |
Seedance 2.5 bills by tokens. The per-second figures are for 16:9 output from a text prompt, from the Seedance 2.5 pricing guide. An agent can call a search API, write a script with a language model and render the video with one balance. Create a key in the AIsa console.
FAQ
Is Veo 3 still available in Google AI Studio?
No. The Veo 3.1 preview models were shut down in the Gemini API on 2026-10-22, and Veo 3 and Veo 3 Fast are deprecated there. Google AI Studio offers Gemini Omni Flash for video generation.
Where can I still use the Veo 3.1 API?
On Google Cloud's Gemini Enterprise Agent Platform, with the GA models veo-3.1-generate-001 and veo-3.1-fast-generate-001 in us-central1. Veo 3.1 Lite is in preview there.
How much does the Veo 3.1 API cost?
$0.40 per second for video with audio at 720p or 1080p, and $0.60 at 4K. Veo 3.1 Fast costs $0.10 to $0.30 per second with audio, depending on resolution.
Is Gemini Omni Flash free?
No. It is on the paid tier of the Gemini API only, at about $0.07 per second of 720p video.
What replaced veo-3.1-generate-preview?
Google names Gemini Omni Flash (gemini-omni-1.2-flash) in the Gemini API, or the GA Veo models on the Gemini Enterprise Agent Platform.
Sources
- Veo, Gemini Omni Flash guide, Gemini Omni Flash model page and pricing, Gemini API docs, read 2026-11-15
- Veo 3.1 model page and Agent Platform pricing, Google Cloud, read 2026-11-15
- The AIsa model catalog, read 2026-11-15