Gemini Omni 1.1 Flash

Positioning Overview

Gemini Omni 1.1 Flash is a multimodal model released by Google for video generation and editing scenarios, and it is called on HiAPI.ai through a unified interface. On a single pipeline, it covers text-to-video, image-to-video, reference-image-to-video, and video editing. It can modify existing clips and extend scenes according to natural-language instructions, and it uses first and last frames to control camera movement and transitions. Output duration ranges from 4 to 10 seconds, with resolutions selectable from 360p to 4k. It is suitable for developers and content teams that need high-frequency draft creation, multi-round iteration, and a unified generation and editing workflow.

Key Highlights

• Generation and editing unified in the same model — text generation, image generation, reference generation, and video editing share a single call path, reducing the cost of switching between multiple models.
• A layered workflow from low-resolution drafts to 4K — you can first use 360p for fast test iterations, and after confirming the direction, produce output at 720p, 1080p, or 4k, balancing efficiency and final quality.
• First and last frames control camera movement — by setting the first-frame and last-frame images, you can constrain the camera direction and transitions, making the motion of narrative clips more controllable.
• Reference materials participate in generation — you can pass up to 7 images as references, and video input and first-frame settings are mutually exclusive, with clear boundaries for input conditions.

Capability Description

Multiple video generation modes: supports four task types: text-to-video, image-to-video, reference-image-to-video, and video-to-video. Developers choose the input method based on the available materials, and the output is uniformly a video asset.

Natural-language video editing: by describing the modification intent in a prompt, you can redraw and adjust an existing video without rebuilding the entire generation pipeline.

Scene extension: by passing an existing video clip as input, you can continue the content on that basis to build a longer continuous narrative.

First and last frame camera constraints: first_frame_url and last_frame_url can be used together to fix the start and end frames and control camera movement and transition pacing.

Selectable resolution and aspect ratio: resolutions support 360p, 720p, 1080p, and 4k, and aspect ratios support 16:9 and 9:16, making it easy to adapt to landscape and portrait placements.

Applicable Scenarios

• Creative teams quickly producing video drafts — input a script or first-frame image and quickly generate multiple drafts at low 360p resolution for script review and direction selection.
• Advertising and social content production — use reference images plus prompts to generate 16:9 or 9:16 clips, and after confirmation, upscale them to high resolution for placement materials.
• Modifying and extending existing materials — take an existing video as input and use natural-language instructions to modify the visuals or extend the clip for post-production adjustments to the final cut.
• Narrative clips that require precise transitions — set first and last frames to constrain the start and end of the camera, producing cinematic clips with controlled motion for storyboard previews.

Limitations and Model Selection Advice

The model's single output duration is limited to 4 to 10 seconds; longer content needs to be generated in segments and then stitched together. 360p results are more suitable as drafts and should not be used directly as the final cut. When a first-frame image is passed, image or video input cannot be added on top of it, and last_frame_url must be used together with first_frame_url, so there are constraints on input combinations. Flash is positioned toward speed and iteration efficiency. If a task has strict image-quality requirements or needs a longer single-segment duration, other models should be evaluated during model selection, and actual test results should be taken as the basis.