> ## Documentation Index
> Fetch the complete documentation index at: https://mulerouter.ai/docs/llms.txt
> Use this file to discover all available pages before exploring further.

# Models cheatsheet

> Browse every model endpoint by capability, with key strengths called out

A curated, hand-grouped index of the endpoints the CLI can run against
MuleRouter. For the canonical machine-readable list run
`mulerouter list --json`. For full parameter detail run
`mulerouter params <endpoint>`.

## Text-to-Image

| Endpoint                     | Highlights                            |
| ---------------------------- | ------------------------------------- |
| `midjourney/diffusion`       | One-prompt, high quality              |
| `alibaba/wan2.6-t2i`         | Batch size, negative prompt, seed     |
| `alibaba/wan2.5-t2i-preview` | Prompt extension                      |
| `google/nano-banana-2`       | Up to 4K, web-search grounding (SOTA) |
| `google/nano-banana-pro`     | Pro tier (SOTA)                       |
| `openai/gpt-image-2`         | 4K, 1–4 images per call (SOTA)        |

## Image-to-Image / Image Editing

| Endpoint                      | Highlights                       |
| ----------------------------- | -------------------------------- |
| `openai/gpt-image-2/edit`     | Mask-based edits, 4K (SOTA)      |
| `google/nano-banana-2/edit`   | Up to 14 reference images (SOTA) |
| `google/nano-banana-pro/edit` | Pro-tier edit (SOTA)             |
| `alibaba/wan2.6-image`        | Image-to-image transform         |
| `alibaba/wan2.5-i2i-preview`  | Preview-tier transform           |

## Text-to-Video

| Endpoint                                    | Highlights                                 |
| ------------------------------------------- | ------------------------------------------ |
| `alibaba/wan2.6-t2v`                        | Multi-resolution, audio, multi-shot (SOTA) |
| `alibaba/wan2.5-t2v-preview`                | Audio support                              |
| `alibaba/wan2.2-t2v-plus`                   | 5-second clips                             |
| `klingai/kling-v3-t2v`                      | Multi-shot, sound (SOTA)                   |
| `klingai/kling-v3-omni-t2v`                 | Omni variant (SOTA)                        |
| `midjourney/video`                          | 480p / 720p (SOTA)                         |
| `bytedance/seedance-2.0-text-to-video`      | Seedance T2V                               |
| `bytedance/seedance-2.0-fast-text-to-video` | Faster Seedance T2V                        |

## Image-to-Video

| Endpoint                     | Highlights                        |
| ---------------------------- | --------------------------------- |
| `alibaba/wan2.6-i2v`         | 1080P, audio (SOTA)               |
| `alibaba/wan2.5-i2v-preview` | Audio support                     |
| `alibaba/wan2.2-i2v-plus`    | 1080P                             |
| `alibaba/wan2.2-i2v-flash`   | Speed-optimized                   |
| `alibaba/wan2.1-kf2v-plus`   | Keyframe interpolation            |
| `klingai/kling-v3-i2v`       | First/last frame, elements (SOTA) |
| `klingai/kling-v3-omni-i2v`  | Omni variant (SOTA)               |

## Video Editing

| Endpoint                         | Highlights                                           |
| -------------------------------- | ---------------------------------------------------- |
| `klingai/kling-v3-omni-v2v`      | Video-to-video (SOTA)                                |
| `klingai/kling-v3-omni-v2v-edit` | Prompted video editing (SOTA)                        |
| `klingai/kling-v3-omni-ref2v`    | Reference-image to video (SOTA)                      |
| `alibaba/wan2.1-vace-plus`       | Outpainting / inpainting / extension / interpolation |

## Discovering more

The list above mirrors the registry baked into the CLI at release time. The
authoritative source is always:

```bash theme={null}
mulerouter list --json
```

To see what changed between CLI releases, check the
[CHANGELOG](https://github.com/openmule/mulerouter-cli/blob/main/CHANGELOG.md).