# Models

Sloosh offers 11 image models and 16 video models. Compare what each is best for, its reference limit, resolution and clip length.

Last updated: 2026-10-05

**Key facts**

- Sloosh offers 11 image models (Nano Banana, GPT Image, FLUX, Seedream, Ideogram and Meta Muse) and 16 video models (Seedance, Kling, Veo, Gemini Omni, MiniMax, Wan, Grok Imagine and Pruna).
- In Sloosh Studio, the GPT Image models take the most reference images, up to 16; the Nano Banana models take 14 and Ideogram 4.5 takes 5.
- Seedance 2.5 makes the longest videos in Sloosh, 4 to 30 seconds, and takes up to 30 reference images.
- Kling 2.5 is the only Sloosh video model with no sound; Veo 3.1, Gemini Omni, MiniMax H3, Grok Imagine and Pruna P-Video 2 Pro always make sound.
- 4K outputs in Sloosh need the Pro or Max plan, on the models that offer 4K.

Web page: https://sloosh.ai/docs/models

## What a model is

A model is the AI that makes your image or video. Each one has its own strengths, settings and price. You pick one in Studio's model chip, and on Spacelab's model nodes. Spacelab also has text and music models; see [Generate nodes](https://sloosh.ai/docs/nodes-generate.md).

Each model below has its own page: when to use it, how to prompt it, a real example with its inputs and prompt, and its settings in Sloosh.

## Image models

| Model | Best for | Max references | Max resolution |
| --- | --- | --- | --- |
| [Nano Banana 2](https://sloosh.ai/docs/nano-banana-2.md) | Fast all-rounder, drafts to 4K | 14 | 4K |
| [Nano Banana Pro](https://sloosh.ai/docs/nano-banana-pro.md) | Finished work, best text | 14 | 4K |
| [Nano Banana Lite](https://sloosh.ai/docs/nano-banana-lite.md) | Cheapest 1K drafts | 14 | 1K |
| [GPT Image 2.5 Sunburst](https://sloosh.ai/docs/gpt-image-2-5-sunburst.md) | Precise edits, fine detail | 16 | 4K |
| [GPT Image 2.5 Flare](https://sloosh.ai/docs/gpt-image-2-5-flare.md) | Fast everyday images with text | 16 | 4K |
| [GPT-image-2](https://sloosh.ai/docs/gpt-image-2.md) | Designed layouts, many instructions | 16 | 4K |
| [FLUX 3 Image](https://sloosh.ai/docs/flux-3-image.md) | Up to 4K, text in many languages | 10 | 4K |
| [Flux 2 Pro](https://sloosh.ai/docs/flux-2-pro.md) | Photoreal scenes, exact hex colours | 8 | 1K |
| [Seedream 5 Pro](https://sloosh.ai/docs/seedream-5-pro.md) | Region edits, multilingual text | 10 | 2K |
| [Ideogram 4.5](https://sloosh.ai/docs/ideogram-4-5.md) | Typography, clean repeat edits | 5 | 2K |
| [Meta Muse](https://sloosh.ai/docs/meta-muse.md) | Concept-rich scenes, world knowledge | 10 | 1K |

## Video models

| Model | Best for | Length | Sound | Inputs |
| --- | --- | --- | --- | --- |
| [Seedance 2.5](https://sloosh.ai/docs/seedance-2-5.md) | Longer stories, timed beats | 4-30 s | Optional | 30 refs, start + end frame |
| [Seedance 2.0](https://sloosh.ai/docs/seedance-2-0.md) | Accurate products, numbered shots | 4, 6, 8, 10, 12 or 15 s | Optional | 9 refs, start + end frame |
| [Seedance 2.0 Fast](https://sloosh.ai/docs/seedance-2-0-fast.md) | Cheaper Seedance 2.0 drafts | 4-15 s | Optional | 9 refs, start + end frame |
| [Kling V3 Omni](https://sloosh.ai/docs/kling-v3-omni.md) | References, frames and multi-shot | 3-15 s | Optional | 7 refs, start + end frame |
| [Kling 3.0](https://sloosh.ai/docs/kling-3-0.md) | Animate a photo with a script | 3-15 s | Optional | Start frame (required) |
| [Kling 2.5](https://sloosh.ai/docs/kling-2-5.md) | Quick silent animation | 5 or 10 s | None | Start frame (required) |
| [Veo 3.1](https://sloosh.ai/docs/veo-3-1.md) | One polished 8-second shot | 8 s | Always | 3 refs, start + end frame |
| [Gemini Omni 1.1 Flash](https://sloosh.ai/docs/gemini-omni-1-1-flash.md) | People, sound, frame-to-frame | 3-10 s | Always | 6 refs, start + end frame |
| [Gemini Omni Flash](https://sloosh.ai/docs/gemini-omni-flash.md) | Older Omni; prefer 1.1 | 3-10 s | Always | 6 refs, start frame |
| [MiniMax H3](https://sloosh.ai/docs/minimax-h3.md) | Ads from text or frames, up to 4K | 5-15 s | Always | Start + end frame |
| [MiniMax H3 Max](https://sloosh.ai/docs/minimax-h3-max.md) | Faster H3, up to 1080p | 5-15 s | Always | Start + end frame |
| [MiniMax H3 Reference](https://sloosh.ai/docs/minimax-h3-reference.md) | H3 guided by references | 5-15 s | Always | 9 refs, start frame |
| [MiniMax H3 Max Reference](https://sloosh.ai/docs/minimax-h3-max-reference.md) | Faster H3 with references | 5-15 s | Always | 9 refs, start frame |
| [Wan 3](https://sloosh.ai/docs/wan-3.md) | Long takes, up to 30 s | 2-30 s | Optional | 10 refs, start + end frame |
| [Grok Imagine](https://sloosh.ai/docs/grok-imagine.md) | Cheap clips from references | 4, 6, 8 or 10 s | Always | 7 refs, start frame |
| [Pruna P-Video 2 Pro](https://sloosh.ai/docs/p-video-2-pro.md) | Fast clips with physics, 768p | 5-15 s | Always | Start + end frame |

## Choosing a model

- Start with the model Studio picks for you.
- Draft on a fast, cheap model, then make the final on a stronger one: Nano Banana Lite then Nano Banana Pro, or Seedance 2.0 Fast then Seedance 2.5.
- Run the same prompt on two models and compare before you commit to a lot of work.
- Check the Generate button when you switch. Models and settings cost different amounts.

## Same words, different models

Most models name references by number: "Image 1" is the first one you add, "Image 2" the second. Veo has no reference tags and matches references by what you describe. Each model's page says how to refer to them.

## What your plan unlocks

Every plan makes up to 4K. Your plan sets the longest video you can make: Pro and Max lift the plan's cap on video length. Each model still has its own longest clip. [Pricing](https://sloosh.ai/pricing) has the details for each plan.

## Frequently asked questions

### What is the best AI image model for product photos in Sloosh?

For finished product photos, start with [Nano Banana Pro](https://sloosh.ai/docs/nano-banana-pro.md), which Sloosh lists for campaigns, e-commerce and product photography. Use [Flux 2 Pro](https://sloosh.ai/docs/flux-2-pro.md) when colours must match exact hex codes, and [GPT Image 2.5 Sunburst](https://sloosh.ai/docs/gpt-image-2-5-sunburst.md) or [Seedream 5 Pro](https://sloosh.ai/docs/seedream-5-pro.md) to change one part of an existing product photo.

### Which AI image model in Sloosh is best for text in images?

Nano Banana Pro has the best text of the Nano Banana models, and Ideogram 4.5 is built for typography such as posters, labels and promo tiles. For layouts with several blocks of text, like menus and recipe cards, try GPT-image-2 or GPT Image 2.5 Flare. Put every word that must appear in quotes.

### Which AI video model in Sloosh makes the longest clips?

Seedance 2.5 makes the longest clips in Sloosh, anywhere from 4 to 30 seconds. Several others go up to 15 seconds, including Seedance 2.0, Kling 3.0, Kling V3 Omni and the MiniMax H3 models, while Veo 3.1 always makes 8 seconds. Your plan can also cap video length.

### Which Sloosh models can make 4K images and videos?

For images, Nano Banana 2, Nano Banana Pro and FLUX 3 Image offer 4K. For video, Veo 3.1, Gemini Omni 1.1 Flash, Kling V3 Omni, Seedance 2.0, MiniMax H3 and MiniMax H3 Reference offer 4K. Every Sloosh plan can make 4K; see [Plans and credits](https://sloosh.ai/docs/plans-and-credits.md).

### What is the cheapest way to test an idea before the final image or video in Sloosh?

Draft on a fast, cheap model, then make the final on a stronger one with the same prompt. In Sloosh that means Nano Banana Lite then Nano Banana Pro for images, or Seedance 2.0 Fast then Seedance 2.5 for video. The Generate button shows the cost before you run.