> For the complete documentation index, see [llms.txt](https://developer.topazlabs.com/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://developer.topazlabs.com/video-models/upscaling-and-enhancement/generative.md).

# Generative Upscaling

Generative Upscaling for Video

* **GenAI video.** Natural faces, skin tones, and stable motion for AI-generated footage.
* **Archival and degraded footage.** Rebuild detail that precision models can only sharpen.
* **Speed vs. quality.** Starlight Precise 2.6 gives the best quality; Starlight Fast 3 is about 4× faster at the same price per frame.
* **Very old, low-quality sources.** Starlight Mini restores VHS, early film transfers, and low-resolution SD with a slight softness that suits archival content. For the best results, run Starlight Mini first, then Starlight Precise 2.6 on the result.

{% hint style="success" %}
We recommend using **Starlight Precise 2.6** for generative upscaling.
{% endhint %}

<table data-view="cards"><thead><tr><th></th><th></th><th data-hidden data-card-cover data-type="image">Cover image</th><th data-hidden data-card-target data-type="content-ref"></th></tr></thead><tbody><tr><td><strong>Starlight Precise 2.6</strong></td><td>Our highest-quality generative upscaler for GenAI and archival video.</td><td><a href="https://4007072434-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FJkSwaYgAsTQke14gpXAF%2Fuploads%2FTzrLBGS7de5x7Zd0EGiN%2Fslp-4k-upscale-example_slider.jpg?alt=media&#x26;token=fd897da0-97a2-4cee-9f9c-bbf1e1da2428">https://4007072434-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FJkSwaYgAsTQke14gpXAF%2Fuploads%2FTzrLBGS7de5x7Zd0EGiN%2Fslp-4k-upscale-example_slider.jpg?alt=media&#x26;token=fd897da0-97a2-4cee-9f9c-bbf1e1da2428</a></td><td><a href="https://developer.topazlabs.com/video-models/upscaling-and-enhancement/generative/starlight-precise-2.6.md">Starlight Precise 2.6</a></td></tr><tr><td><strong>Starlight Fast 3</strong></td><td>Our fastest diffusion upscaler, about 4× faster than Precise 2.6.</td><td><a href="https://4007072434-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FJkSwaYgAsTQke14gpXAF%2Fuploads%2FWxbaubRraS4eiZeGo6Ws%2Fvideo19-seedance1.5pro-slf3_slider.jpg?alt=media&#x26;token=29661a0b-3eb0-482d-9ac9-68ba0f89d114">https://4007072434-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FJkSwaYgAsTQke14gpXAF%2Fuploads%2FWxbaubRraS4eiZeGo6Ws%2Fvideo19-seedance1.5pro-slf3_slider.jpg?alt=media&#x26;token=29661a0b-3eb0-482d-9ac9-68ba0f89d114</a></td><td><a href="https://developer.topazlabs.com/video-models/upscaling-and-enhancement/generative/starlight-fast-3.md">Starlight Fast 3</a></td></tr><tr><td><strong>Starlight Mini</strong></td><td>Diffusion restoration for very old, low-quality, and interlaced footage.</td><td><a href="https://4007072434-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FJkSwaYgAsTQke14gpXAF%2Fuploads%2Faafy3V3Gk6RBKPgpJraZ%2FStarlight_Mini_(Ex_2)_slider.jpg?alt=media&#x26;token=fd4719f0-c78a-4c7e-a0bf-0c31191bd53e">https://4007072434-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FJkSwaYgAsTQke14gpXAF%2Fuploads%2Faafy3V3Gk6RBKPgpJraZ%2FStarlight_Mini_(Ex_2)_slider.jpg?alt=media&#x26;token=fd4719f0-c78a-4c7e-a0bf-0c31191bd53e</a></td><td><a href="https://developer.topazlabs.com/video-models/upscaling-and-enhancement/generative/starlight-mini.md">Starlight Mini</a></td></tr><tr><td><strong>Starlight Fast 2</strong></td><td>Lower-cost diffusion upscaling for price-sensitive workloads.</td><td><a href="https://4007072434-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FJkSwaYgAsTQke14gpXAF%2Fuploads%2FtJ6UpzLD7BqYlDHmG2Xi%2FStarlight_Fast_2_slider.jpg?alt=media&#x26;token=fb24656a-3977-46d4-a8bb-e2e95de7ed9e">https://4007072434-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FJkSwaYgAsTQke14gpXAF%2Fuploads%2FtJ6UpzLD7BqYlDHmG2Xi%2FStarlight_Fast_2_slider.jpg?alt=media&#x26;token=fb24656a-3977-46d4-a8bb-e2e95de7ed9e</a></td><td><a href="https://developer.topazlabs.com/video-models/upscaling-and-enhancement/generative/starlight-fast-2.md">Starlight Fast 2</a></td></tr></tbody></table>

{% hint style="info" %}
**Starlight Fast 2** is half the price of Starlight Fast 3. Starlight Fast 3 is substantially faster and higher quality, so we recommend it for all new integrations.
{% endhint %}

{% hint style="warning" %}
Starlight Precise 1, Starlight Precise 2, Starlight Precise 2.5, Starlight HQ, Starlight Sharp, and Starlight Fast 1 are legacy models and are no longer recommended. Starlight Precise 2.6 replaces Starlight Precise 1, Starlight Precise 2, Starlight. Precise 2.5, Starlight HQ, and Starlight Sharp. Starlight Fast 3 replaces Starlight Fast 1. These models will be retired no sooner than December 9, 2026. See [Model Deprecations](https://developer.topazlabs.com/resources/model-deprecations.md) for dates and migration steps.
{% endhint %}

### Pricing

Generative models are billed per output frame:

```
credits = ceil(output frames ÷ frames per credit)
```

| Model                                                   | 1080p                 | 4K                    |
| ------------------------------------------------------- | --------------------- | --------------------- |
| Starlight Precise 2.6, Starlight Fast 3, Starlight Mini | 26.0417 frames/credit | 11.9179 frames/credit |
| Starlight Fast 2                                        | 52.0833 frames/credit | 23.8357 frames/credit |

{% hint style="info" %}
See 'Per Output Frame' in [Video Pricing](https://developer.topazlabs.com/getting-started/model-pricing/video-models.md) for a full breakdown, or use the [Credit Calculator](https://developer.topazlabs.com/credit-calculator) for a quote on a specific file.
{% endhint %}


---

# 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 by asking a question.

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

```
GET https://developer.topazlabs.com/video-models/upscaling-and-enhancement/generative.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 `automate deployments from our CI pipeline` 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.
