# AI image and video models

> Compare available image and video models by workflow, supported inputs, output controls, and credit cost in the editor. This directory does not provide a measured quality or speed ranking.

Canonical URL: https://imaginescenes.com/models

## Image models

### GPT Image 2.5 Flare

Generate an image from a written brief or edit an uploaded reference. Choose the exposed size and output controls, then review faces, text, and product details.

### GPT Image 2.5 Sunburst

Explore another GPT Image 2.5 image-generation or editing direction. Compare the supported settings, credit estimate, and result with your brief before choosing a version.

### GPT Image-2

Create prompt-led images and reference-guided edits. Quote short text explicitly and inspect spelling, layout, and small details in the downloaded result.

### Nano Banana 2

Generate or edit images with the controls and reference slots exposed in the editor. Compare the source and result for identity, labels, and composition.

### Nano Banana Pro

Use a written brief or supported image references for image creation and editing. Assign reference roles and review the available output settings and cost.

### Nano Banana 2 Lite

Explore an image concept or edit with this Nano Banana variant. Check its current upload and output options; compare the estimate before submitting another version.

## Video models

### Seedance 2.5

Create a video from text or compatible visual inputs. Describe one action and assign distinct reference roles; check duration options and the live cost in the selected workflow.

### Seedance 2.0

Create a scene from text or supported reference media. Use the selected workflow to review upload slots, output choices, and cost; inspect the full generated clip.

### Seedance 2.0 Fast

Explore a Seedance scene brief with this variant. Compare its exposed settings and cost with other variants; the tier name does not guarantee turnaround or result quality.

### Seedance 2.0 Mini

Test a focused Seedance scene direction with compatible inputs. Choose from the options shown in the editor and review the estimate and resulting motion.

### MiniMax H3

Create a video with text or compatible reference inputs. Name the action and camera direction, check the available output options, and review subject detail across the clip.

### Wan 3.0

Use a supported text, image, or footage workflow to guide a new video. Input types and output controls depend on the mode; check the editor before submission.

### Gemini Omni 1.1 Flash

Guide a Gemini video with text or compatible source media. Choose the workflow first, review its frame and reference controls, and inspect transitions in the result.

### Gemini Omni

Guide a new Gemini video with the reference inputs available in the selected workflow. Assign each source a purpose and check its upload limits and credit estimate.

### Veo 3.1 Quality

Create a Veo scene from text or compatible image inputs. Check frame, reference, and sound options in the selected workflow, then review the complete clip and audio.

### Veo 3.1 Fast

Explore a Veo scene with this variant and its available output controls. Compare settings and cost before generating; review motion and any generated sound.

### Veo 3.1 Lite

Create a Veo video with the options exposed for this variant. Check supported input mode, duration, framing, and sound before comparing the estimate.

### Kling 3.0

Guide a Kling clip with text or supported frame inputs. Describe one continuous action and inspect the opening, transitions, and any generated sound.

### Kling 3.0 Omni

Use the supported frame, reference, or video-transformation mode to guide a Kling scene. Check upload slots and output options; exact identity and timing can vary.

### Kling 3.0 Turbo

Explore a Kling clip from text or compatible image input. Check the available length, resolution, and cost, then inspect motion and subject detail throughout.

## Capability boundaries

- Available models, inputs, output settings, and credit costs vary by workflow and provider.
- Generated results depend on the selected model and prompt; no particular result is guaranteed.
- Use the HTML page for the interactive creation workflow and its current controls.

Updated: 2026-10-01