Gemini API moves Omni 1.1 Flash calls to 1.2 from 3 December

Gemini API moves Omni 1.1 Flash calls to 1.2 from 3 December

5 November 2026: Google will automatically route every Gemini API request for gemini-omni-1.1-flash to gemini-omni-1.2-flash from 3 December, at the newer model's rates. The newer model cuts video output pricing by 30%, so the bill should fall, but an application's video generation will change model on that date with no change on the developer's side.

Not yet independently verified. Dates and prices are from Google's own changelog and pricing page. We have not found a Google statement on whether output quality or behaviour changes between the two versions. We will update this when it can be confirmed, and remove this note.

The price change, in tokens and in seconds

Omni 1.2 Flash, now generally available, charges $12 per million tokens for video output, against $17.50 on 1.1. Input stays at $1.50 per million tokens, and text output at $9. Neither model has a free tier.

Google bills 720p video at 5,792 output tokens per second. In clip terms:

• Omni 1.2 Flash at 720p: about $0.07 a second, so about $0.70 for 10 seconds and about $1.04 for 15;
• Omni 1.1 Flash: about $0.10 a second;
• Omni 1.2 Flash at 360p: $0.023 a second;
• Omni 1.2 Flash at 1080p and 4K, with upsampling: $0.104 and $0.209 a second.

The higher tiers matter when budgeting. A team that moves to 4K because it is now on offer pays three times the 720p rate, so the 30% cut could easily be cancelled by a change in default settings.

What else arrives with 1.2

Clips can run up to 15 seconds with native audio, with conversational editing between turns. Each turn can extend a clip by 3 to 10 seconds, up to 30 seconds in total. First and last frame interpolation covers 3 to 10 seconds, and up to six reference images keep a character or subject consistent. Google has also doubled paid-tier rate limits.

An earlier preview model, gemini-omni-flash-preview, has been shut down. Code still pointing at it needs attention now rather than in December.

Testing before the switch

There is nothing to do to receive the routing, which is the reason to act before it. A product that generates customer-facing video has prompts, duration settings and perhaps a human review step tuned to what 1.1 produces. Pointing a test environment at gemini-omni-1.2-flash ahead of time shows what will change while the old version is still there as a comparison, and any cost tracker keyed on the model name needs updating as well.

Hard-coding gemini-omni-1.2-flash also makes the choice explicit, and leaves a clear record of which model each clip came from.

Our profile of Gemini Omni 1.2 Flash sets out the specification and limits. Gemini Omni 1.1 Flash has its own entry until the routing date.

Sources

• Gemini API changelog, entry for 5 November 2026, 5 November 2026
• Gemini API pricing (read 6 November 2026), 6 November 2026
• Gemini Omni Flash model page (read 6 November 2026), 6 November 2026
• OpenRouter: Gemini Omni 1.2 Flash, released 5 November 2026, 5 November 2026