AI image models built intoyour video workflow
Google's Nano Banana 2 and OpenAI's GPT Image 2.5 generate the stills your scenes, B-roll and thumbnails need — then hand them straight to the timeline.

Nano Banana 2
Google · Image
Google's image model for photographic scene stills, at 1K, 2K or 4K.

Nano Banana 2 Lite
Google · Image
Google's fast, low-cost 1K image tier for high-volume AI B-roll.

GPT Image 2.5
OpenAI · Image
OpenAI's image model (Flare) — legible text and fast edits, the thumbnail default.

GPT Image 2.5 Sunburst
OpenAI · Image
OpenAI's precision variant — sharper detail, natural light, strongest same-face likeness.
Compare image models
Credit prices are read live from the same catalog the app bills from (1 credit = €0.01).
| Model | Best at | Max size | Speed | Credits |
|---|---|---|---|---|
| Google's image model for photographic scene stills, at 1K, 2K or 4K. | 4K | Standard | 6–20 per image | |
| Google's fast, low-cost 1K image tier for high-volume AI B-roll. | 1K | Fast | 6–13 per image | |
| OpenAI's image model (Flare) — legible text and fast edits, the thumbnail default. | 1536px | Fast | 1–28 per image | |
| OpenAI's precision variant — sharper detail, natural light, strongest same-face likeness. | 1536px | Slower | 1–28 per image |
Every model, one app
Shorz isn't a model playground. The models generate the parts; the editor around them turns those parts into a finished, captioned, published video.
- Every model on one prepaid balance — no per-provider subscriptions
- No watermark on anything you generate or export
- Full commercial rights on what you create
- No API keys to manage: the models run on your Shorz account
- New model versions appear in the picker without an app update
- Drivable from Claude, Cursor, Codex or Gemini over MCP
AI image models questions
Nano Banana 2 is the default for scenes and B-roll and the only one with 2K and 4K output. Nano Banana 2 Lite is the cheap, fast tier for B-roll. GPT Image 2.5 is the default for thumbnails and the pick when an image has to carry legible text; GPT Image 2.5 Sunburst trades a few seconds for sharper detail and the strongest same-face likeness.
Yes. Pass reference images — a face photo works best — and the model keeps that identity in every generation, so scene 12 still looks like scene 1.
A few credits each (1 credit = €0.01). The exact per-image price for every model and size is on each model page and comes from the same live catalog the app bills from.
Still have questions? Ask founder
More model families
AI video models
Text-to-video and image-to-video clips for scenes, B-roll and ads.
AI avatar models
Lip-synced talking presenters from one photo and a voice track.
AI language models
The brain of every project: scripts, edit plans and the agent that runs them.
AI voice & audio models
Narration, transcription, dubbing, sound effects and voice cleanup.
AI music models
Royalty-free background tracks generated from a description.
Every image model, one picker
Download Shorz for Windows or Mac — every model on this page is already installed and billed from one prepaid balance.
Publishes straight to

