How does pricing work?

Runware pricing is based on the time it takes to generate an image for models that are natively hosted on Runware.

Pricing is per image and will vary based on the time taken to generate it. This will vary depending on the models, technologies, or workflows you choose. With thousands of models and configurations used, we don’t offer a fixed price list for each.

How pricing is calculated

The pricing principle is straightforward:

If an image takes twice as long to generate as the reference, it will cost twice as much.

This means pricing scales directly with the compute time required for the generation.

What affects the price of an image?

The cost of generating an image can vary depending on:

• The model used

• The technologies or workflow involved

• The time required for the generation

Since Runware supports thousands of models and configurations, pricing is based on actual generation time rather than a fixed price per model.

Pricing for third-party models

Some models available on Runware are provided by third-party model providers.

For these models, pricing is determined by the provider, and the cost may differ from the native Runware pricing model. The provider’s pricing and rules will apply to those requests.

Why Runware uses this pricing model

Runware is built for speed. Thanks to our market-leading inference times, we can generate images extremely quickly and pass our cost savings directly to users.

This approach ensures users only pay for the actual resources used during generation.