After Sora's shutdown: 5 product teasers on Omni Flash, $5.00

If your product teasers ran on sora-2, you need a new route: OpenAI's deprecations page says the Videos API and the sora-2 family were removed on 2026-09-24, with no replacement listed. On Sume the default for video is Gemini Omni Flash 1.1, and five 8-second 720p teasers cost 5 x 100 cents = $5.00 (read 2026-10-08).

What OpenAI's page says

The OpenAI deprecations page (read 2026-10-08) lists the Videos API plus sora-2, sora-2-pro, sora-2-2025-10-06, sora-2-2025-12-08 and sora-2-pro-2025-10-06. Developers were notified on 2026-03-24, the removal date is 2026-09-24, and the page names no replacement. Any job that still calls those ids has to move.

• Notified: 2026-03-24
• Removed: 2026-09-24
• Replacement listed by OpenAI: none

The Sume route and what it costs

Sume's POST /v1/videos follows the OpenRouter-style flow: submit, poll the polling_url, then download unsigned_urls[0]. With model: "sume/auto" Sume picks the family, and today that is Gemini Omni Flash 1.1. It makes 3 to 10 second clips at 360p, 720p, 1080p or 4K, in 16:9 or 9:16 only, with native synced audio always on. Pin the model by id if you need a stable choice.

Port the recipe, not just the call

A teaser is a prompt, a duration and an aspect ratio. Keep those three in your own repo, then re-run them against the new route. Three differences matter in practice. Omni Flash gives no square or 21:9, so a square placement needs a crop afterward. Audio is always generated, so you cannot ask for a silent clip. And the clip tops out at 10 seconds, so a longer teaser is two clips joined on Timeline 1.0 at $0.10 per output minute.

curl -X POST https://api.sume.com/v1/videos \
  -H "Authorization: Bearer $SUME_API_KEY" \
  -H "Content-Type: application/json" \
  -H "Idempotency-Key: teaser-01" \
  -d '{
    "model": "sume/auto",
    "prompt": "Product on a desk, slow push-in, soft daylight, small logo in frame",
    "resolution": "720p",
    "duration": 8,
    "aspect_ratio": "9:16"
  }'