Veo 3.1 preview API shuts down Oct 22: what to switch to

Heads up if you call Veo 3.1 from code. Google's Gemini API docs mark all three Veo 3.1 preview model IDs as deprecated with a shutdown date of October 22, 2026: veo-3.1-generate-preview, veo-3.1-fast-generate-preview and veo-3.1-lite-generate-preview. The Veo guide now opens with a banner saying the same thing. Once a model is shut down, Google says, the endpoint is no longer available, so any script, agent skill or pipeline that hard-codes one of these IDs will start failing.

What to switch to: the deprecations table lists gemini-omni-1.1-flash (Gemini Omni Flash, released Aug 27, 2026, no shutdown date announced) as the replacement for all three. The Veo guide's banner adds a second route: the GA models available through the Gemini Enterprise Agent Platform. Two traps to avoid. Don't fall back to gemini-omni-flash-preview, which shuts down on the same day. And don't fall back to the older veo-3.0-generate-001 or veo-3.0-fast-generate-001 IDs either, which the same table lists with a June 30, 2026 shutdown date.

What this does not mean: this notice covers the Gemini API preview model IDs only. Google's pages don't say Veo 3.1 is going away in Flow, on Vertex AI's GA models or at third-party hosts, and we haven't checked each of those, so treat them as unaffected until their own docs say otherwise. Our model picker still lists Veo 3.1 for that reason.

Practical steps: search your repos, skills and n8n or ComfyUI configs for 'veo-3.1-' and '-generate-preview'. Swap in the new model ID and re-run a few of your known-good prompts before trusting batch jobs, because we haven't tested whether Omni Flash matches Veo 3.1 on clip length, 4K output, reference images or extension. Spotted, not tested. Source: Google's Gemini API docs, checked Oct 21, 2026.