
Google has pushed Gemini Omni 1.1 Flash away from the clever clip machine and towards something resembling an edit bench. Released on 27 August, the model can inspect up to 10 seconds of preceding footage, rather than only the final second, then extend a sequence in 10-second steps to a cumulative 40 seconds. That is not a feature film, but it may stop every continuation feeling as if continuity arrived by fax.
Google DeepMind Watch analysis
More control between the cut points
Developers can supply both a first and last frame and ask Omni to generate the motion between them. Outputs run for three to 10 seconds at 24fps. The default is 720p; 1080p and 4K are upscaled outputs, an important distinction from native high-resolution generation.
There is also a 360p drafting mode. Google says it is up to 60 per cent faster and costs one-third as much as 720p. That is a company comparison, not an independent benchmark, but the workflow makes sense: audition cheap roughs, then spend on the take that survives contact with human taste.
The cheaper scout and the dearer finish
There is no free Gemini API tier for this model. Google’s launch table lists output at $0.03 per second for 360p, $0.10 for 720p, $0.15 for 1080p and $0.30 for 4K. At those rates, a maximum 10-second render costs $0.30, $1.00, $1.50 or $3.00 respectively, before input charges. The API pricing page separately lists input at $1.50 per million tokens and video output at $17.50 per million, with 720p working out at roughly $0.10 per second.
GA, Preview and the small type
The stable gemini-omni-1.1-flash endpoint is generally available on the paid Gemini API. The older gemini-omni-flash-preview endpoint is due to retire on 30 September 2026, while Google’s Enterprise Agent Platform separately labels gemini-omni-1.1-flash-preview as Preview. AI Studio, Flow and Enterprise are in the rollout; Google also says scene extension is available globally in the Gemini app for AI Plus, Pro and Ultra subscribers. One launch, several labels: procurement teams may wish to bring a highlighter.
- Uploaded video is capped at 10 seconds for editing or extension.
- Google’s model card says complete edit consistency, complex motion and perfectly accurate text remain difficult.
- Longer context is a continuity aid, not evidence that identities, lighting or physics will stay locked on every take.
Sources and evidence
- Gemini Omni 1.1 Flash lets you build with more control (27 August 2026, 16:11 UTC)
- Gemini Omni Flash | Gemini API (27 August 2026)
- Gemini Developer API pricing (28 August 2026)
- Gemini API release notes (27 August 2026)
- Gemini Omni 1.1 Flash Preview | Gemini Enterprise Agent Platform (27 August 2026)
- Gemini Omni Flash – Model Card (27 August 2026)
- Google launch thread for Gemini Omni 1.1 Flash (27 August 2026, 16:11 UTC)
Google DeepMind Watch is an independent WittyWires-operated tracker and is not affiliated with, endorsed by, or operated by Google DeepMind.