Stručne o hlavnom (BLUF): Gemini Omni 1.2 Flash is the new version of Google's multimodal video model. A clip is built from a prompt and up to seven reference images, duration is set to any second between 3 and 15, quality runs from 360p drafts to 4K, and audio is generated together with the video. The main gain over 1.1 Flash is twice as long clips and a fine-grained duration scale.
Three changes over version 1.1. First, duration: a clip can now run up to 15 seconds instead of 8–10. Second, the duration scale: you pick any whole second from 3 to 15, so the runtime fits the task instead of a preset. Third, the reference mode is reworked: the first uploaded image can be set as the starting frame, the rest act as style hints.
Everything else is familiar Omni workflow: a 360p draft mode for quick idea runs, 720p, 1080p and 4K for finals, landscape 16:9 and portrait 9:16 frames, and a seed field for repeatable results.
Duration is chosen as an exact number of seconds: 5 seconds for a story clip, 12 for a dialogue scene. The runtime is trimmed to the task, not to a four-step ruler.
Pricing is per second: the longer the clip and the higher the resolution, the higher the total. A 360p draft costs noticeably less than a 1080p or 4K final, so the practical route stays the same — iterate the composition in 360p, then reshoot the same prompt with the same seed in high resolution.
A request accepts up to seven images. By default each one works as a reference: the model reads style, palette and scene composition from them. The “first image is the starting frame” checkbox turns the first file into the literal first frame — motion starts from your picture, and the rest of the scene is described in words.
Video and audio inputs, preset voices and saved characters are not available in 1.2 Flash: this is a text-plus-images mode. If a scene needs a voice or a video reference, the neighbouring Gemini Omni versions in the same model list provide them.
Audio is generated together with the video: the model voices the scene itself, no separate track to upload. The seed field fixes the random part — the same prompt with the same seed repeats the take's composition, differing only in detail between resolutions.
The generation history is shared with the other Gemini Omni versions: 1.2 Flash clips live in the same feed, and any good take can be published as a trend or opened in the editor.
Pick 1.2 Flash when a clip must run longer than ten seconds or the runtime must be exact to the second. 1.1 Flash remains the choice where voices, characters and video input matter — 1.2 lacks those modes but offers a finer duration scale.
Switching between versions breaks nothing: form settings are shared, the history is shared, and pricing follows the same per-second rules. Both versions live in one group of the Video list.
Duration and its scale: 1.2 shoots clips at any second from 3 to 15 (1.1 tops out at 8–10), while 1.2 has no voices, characters or video input. Frame formats, 360p–4K, seed and audio are shared.
Yes. Pricing is per second and depends on resolution: a 360p draft costs noticeably less than 1080p and 4K for the same seconds, so drafts are best shot in 360p and finals in high resolution.
No. 1.2 Flash takes a prompt and up to seven images; the first one can be set as the starting frame. Video input, preset voices and saved characters are available in the neighbouring Gemini Omni versions.
Yes. Fix the seed of a good draft, repeat the prompt and choose 1080p or 4K — the composition stays, only the detail grows.