Runchat installs several popular models from Fal when you connect your API key. These models offer a spectrum of choice between fast cheap and low quality models, and higher quality slower and more expensive options.
Runchat currently offers two models to users for free.
| Model | Description | Max Size (px) | Inference Steps | Image to Image? | Vector output |
|---|---|---|---|---|---|
| fast-turbo-diffusion | Ultra fast low quality model | 768 | 2 | ||
| fast-turbo-diffusion/image-to-image | Ultra fast low quality model | 768 | 5 | ✅ |
The default models are installed when you connect a Fal API key.
| Model | Description | Max Size (px) | Inference Steps | Image to Image? | Vector output |
|---|---|---|---|---|---|
| flux/schnell | Fast medium quality model without image to image support | 1536 | default | ||
| stable-diffusion-v35-large/turbo | Fast medium quality model without image to image support | 1536 | default | ||
| flux/dev | Slower high quality model | 1536 | 20 | ||
| flux/dev/image-to-image | Slower high quality model | 1536 | 40 | ✅ | |
| flux-pro/v1.1 | Highest quality model without image to image support | 1536 | default | ||
| fal-ai/recraft-v3 | High quality realistic images capable of vector illustration | 1536 | default | ✅ |
Fal provides dozens of endpoints for controlling image generation, working with other, newer models, generating video and audio and more. Check out their API docs for more information: