Veo 3.1 after the Gemini API shutdown (October 2026)

Updated October 31, 2026, 18:00 UTC. UberLLM prices are read live when the page is built (every hour); Google’s notice and other providers checked 2026-10-30.

Google shut down Veo 3.1 Preview on the Gemini API on October 22, 2026; calls to veo-3.1-generate-preview and the Fast and Lite previews no longer work. On October 31, 2026, UberLLM sold Veo 3.1 Fast with sound at $0.12 per second, $0.96 for an 8-second clip, through one OpenAI-style video endpoint. Google points users to Gemini Omni Flash or its GA Veo models on Vertex AI, which are not sold pay-as-you-go. kie.ai sells the same clip for $0.30.

“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. Migrate to Gemini Omni Flash (gemini-omni-1.1-flash) or the GA models available through the Gemini Enterprise Agent Platform to avoid service interruptions.”

Your options

• Gemini Omni Flash on the Gemini API, Google’s suggested replacement: a different model, so prompts and output change.
• Veo 3.1 GA on Vertex AI (veo-3.1-generate-001, veo-3.1-fast-generate-001): the same model family under Google’s enterprise terms, but its model page lists only Provisioned Throughput and fixed quota, not pay-as-you-go.
• A Veo 3.1 API from a reseller, pay per video: UberLLM, kie.ai, APIMart, fal, Replicate or WaveSpeedAI. Prices for each are on Cheapest Veo 3 API.

Veo 3.1 on UberLLM now

Read October 31, 2026. A job costs the seconds times the rate for what you ask; the reply says the most it can cost before it runs. kie.ai, read from its public price list at build: Veo 3.1 Fast 720p $0.30 per video, Quality 720p $1.25, the same for 4, 6 or 8 seconds. Video API pricing · Fal video pricing

The same request, before and after

Before: the Gemini API (google-genai SDK), which now fails.

from google import genai
import time

client = genai.Client()
operation = client.models.generate_videos(
    model="veo-3.1-generate-preview",  # shut down October 22, 2026
    prompt="a red ball bouncing on green grass",
)
while not operation.done:
    time.sleep(10)
    operation = client.operations.get(operation)
video = operation.response.generated_videos[0]
client.files.download(file=video.video)
video.video.save("video.mp4")

After: UberLLM. Same prompt; start the job, check on it, download the file.

import os, time, requests

API = "https://api.uberllm.dev"
H = {"Authorization": f"Bearer {os.environ['UBERLLM_API_KEY']}"}

job = requests.post(f"{API}/v1/video/generations", headers=H, json={
    "model": "fal-ai/veo3.1/fast",
    "prompt": "a red ball bouncing on green grass",
    "duration_seconds": 8,   # 4, 6 or 8
    "resolution": "720p",    # 720p, 1080p or 4k
    "audio": True,           # Veo adds sound unless you send False
}).json()
print("most this job can cost:", job["max_cost_usd"])

while True:
    r = requests.get(f"{API}{job['poll_url']}", headers=H)
    if r.status_code == 200:  # 202 while it runs
        break
    time.sleep(8)
done = r.json()
if done["status"] != "succeeded":
    raise SystemExit(done)  # failed jobs are not billed
video = requests.get(done["results"][0]["url"], headers=H)
open("video.mp4", "wb").write(video.content)

Full request and reply fields: images, video, music and speech. No UberLLM fee on requests. Adding credit by card costs 5.5% + $0.50.

Questions