> For the complete documentation index, see [llms.txt](https://help.intangible.ai/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://help.intangible.ai/visualize/resolution-and-cost.md).

# Resolution and cost

1K, 2K and 4K, what each step actually costs on each model, and when the higher resolution is worth it.

Image generation runs at 1K, 2K or 4K. Video has its own ladder, further down this page.

Resolution and model both drive what a render costs. Higher resolutions cost more, and the newer, more capable models cost more than the older ones.

![The resolution pill reading 2K with its Output Quality list open below it, offering 1K, 2K and 4K with 2K ticked](https://1179478100-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FSWrjIqlgVAblx6vHGJiM%2Fuploads%2Fgit-blob-129a2bf2e4f8ef24b0747e17cdf6d3e470e5fc0d%2Fresolution-toggle-01.png?alt=media)

The control is the small pill beside the model name at the bottom of the panel. Clicking it opens a short list headed **Output Quality**, which drops downward rather than opening upward. The rungs it offers depend on the model.

## How the steps compare

There is no single multiplier. The gap between 1K and 2K is small on some models and substantial on others, and 4K is a real step up everywhere it is offered. Which model you are on changes the answer enough that a rule of thumb will mislead you.

So do not budget from a ratio. Run a couple of renders at the resolution you are considering and watch the credit counter move. That is the only number that reflects your model, your settings and your plan.

{% hint style="warning" %}
The generate button does not quote a price before you click. Video is the exception: an expensive one warns you first that it is a large generation, and tells you how many credits it will consume.
{% endhint %}

## When 4K is worth it

4K costs meaningfully more than 1K on every model that offers it, so it earns its place in fewer situations than 2K does:

* **Print and large-format display.** The one case where the pixels are the point.
* **The hero shot of a sequence.** When a single image carries the story.
* **Renders you will crop into.** 4K gives you room to reframe in post without softening.
* **Anything with fine text or small hardware detail** that survives at 4K and mushes at 1K.

Not for iteration, and not for a first look. 2K tells you whether you like it for a fraction of what 4K costs.

## What you actually get

Pixel dimensions depend on the aspect ratio, and the numbers are not always the round ones you would expect. Both edges have to divide by 16 and the total pixel count is bounded, so:

| Aspect | 1K          | 2K          | 4K          |
| ------ | ----------- | ----------- | ----------- |
| 16:9   | 1280 x 720  | 1920 x 1088 | 3840 x 2160 |
| 1:1    | 1024 x 1024 | 2048 x 2048 | 2880 x 2880 |

**2K 16:9 is 1920x1088, not 1920x1080**, because 1080 is not a multiple of 16. And 4K square tops out at 2880, not 4096, because the square runs into the pixel ceiling first. If you are delivering to an exact spec, check the number rather than assuming the label.

## Video resolution and cost

Video generation has its own resolution scale that varies per model provider. The resolution and orientation dropdowns filter as you pick a model; the visualizer surfaces only the options that model supports. Any model offering more than one resolution shows the selector, so the options change as you move between models.

![The resolution selector open on a video shot, Seedance 2.5 selected, offering 720p and 480p](https://1179478100-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FSWrjIqlgVAblx6vHGJiM%2Fuploads%2Fgit-blob-c3a06581b614d497bf2c4f69c70e30b9160d3a4c%2Fvideo-resolution-selector-01.png?alt=media)

The ladder runs 480p, 720p, 1080p, 4K, with one extra rung: **H3 Max renders at 768p**, its own native tier, and that option appears only while H3 Max is the selected model. Most models in the working set cover 720p and 1080p. Pricing is per resolution, so a 4K render costs materially more than the same shot at 720p.

Four paths reach 4K:

| Path             | Notes                                                                                                                         |
| ---------------- | ----------------------------------------------------------------------------------------------------------------------------- |
| **Kling 3 (4K)** | A separate model on its own endpoint, picked from the model dropdown rather than the resolution selector.                     |
| **Seedance 2.0** | Covers the full ladder, 720p through 4K.                                                                                      |
| **Omni Flash**   | 1080p and 4K are upscales of the same generation rather than native renders, and pricing scales with the resolution you pick. |
| **FLUX.3**       | Added in 2.0.72. Labels its rungs 1K, 2K and 4K rather than in p, the same scale the image models use.                        |

Two models stop short of 1080p: **Seedance 2.5** and **Seedance 2.0 Mini** run at 480p and 720p only. Picking either for a 1080p deliverable means regenerating on a different model.

**H3 Max** reaches 1080p but not 4K, and its three rungs are 480p, 768p and 1080p. Added in release 2.0.73.

{% hint style="warning" %}
**Luma Ray 3.14 is not a 4K path.** It carries 720p and 1080p. If you need 4K from an authored 3D animation, Seedance 2.0 is the closest equivalent.
{% endhint %}

Veo 3.1 is the one model where the two lower rungs cost the same: 720p and 1080p carry identical pricing, and 720p is the default. If the deliverable is going to 1080p, set it before you generate rather than re-rendering afterwards.

Audio toggle is independent of resolution and adds cost on top. If you don't need audio, leave it off – audio generation can roughly double a video render's cost.

{% hint style="info" %}
Resolution is worth deciding before the render, not after. There's no upscale step inside the visualizer, so a 720p draft you decide to deliver has to be generated again at the higher resolution and will not come back identical.
{% endhint %}

## Cost as a project budget signal

A common pattern across projects:

```mermaid
flowchart LR
    A[Concept exploration] --> B[Iterate at 1K or 2K, many models]
    B --> C[Composition locked]
    C --> D[Iterate, fewer models, more presets]
    D --> E[Look locked]
    E --> F[4K hero renders for delivery]
    E --> G[2K supporting renders]
```

The shape of a project's spend is iteration at a lower resolution, then a small number of high-resolution renders for delivery. Most of the budget goes on the iteration, and that is the right shape.

If a project is burning credits faster than expected, the usual cause is high-resolution renders too early, before the composition and the look are settled. Drop back down until they are.

## Limits and known issues

* **Resolution doesn't compensate for the wrong model.** A 4K render with the wrong model is worse than a 1K render with the right model. Pick the model first, then the resolution.
* **Pricing varies by model provider.** A 1K render on one model can cost more than a 4K render on another, so the ratios above only compare a model against itself. The dropdown reflects what's available; pick the model that fits the shot, then pick the resolution.
* **4K is not offered by every model.** Seedream 5 Pro stops at 2K, and Flux2 Pro runs on a different scale entirely. If 4K is a delivery requirement, it constrains which model you can use.
* **Video resolution affects generation time more than image resolution does.** Higher-resolution video renders can take noticeably longer; plan accordingly when the deadline is short.

## Related

* [Generate image](/visualize/generate-image.md)
* [Generate video](/visualize/generate-video.md)
* [Credits and tokens](/overview/concepts/credits-and-tokens.md)
* [Models](/visualize/ai-models.md)
* [Plans and billing](/teams-and-billing/plans-and-billing.md)


---

# Agent Instructions
This documentation is published with GitBook. GitBook is the documentation platform designed so that both humans and AI agents can read, navigate, and reason over technical content effectively. Learn more at gitbook.com.

## Querying This Documentation
If you need additional information that is not directly available in this page, you can query the documentation dynamically by asking a question.

Perform an HTTP GET request on the following URL with the `ask` and `goal` query parameters:

```
GET https://help.intangible.ai/visualize/resolution-and-cost.md?ask=<question>&goal=<user_goal>
```

`ask` is the immediate question: it should be specific, self-contained, and written in natural language.
`goal` is what the user is ultimately trying to achieve, the reason they need the answer. Sharing it helps GitBook give you a better, more relevant answer. A goal is most helpful when it describes the outcome the user wants rather than restating the question. For example, with `ask=how do I create an API token`, a goal like `build a script that syncs our docs to a CMS` lets GitBook tailor the answer to that use case.

The response will contain a direct answer to the question and relevant excerpts and sources from the documentation.

Use this mechanism when the answer is not explicitly present in the current page, you need clarification or additional context, or you want to retrieve related documentation sections.