Wan3.0 Prime
Alibaba's bigger video model, same node-shaped door
- referenceImages
- frameImages
- referenceVideos
- video
Wan3.0 Prime is the step-up tier of Alibaba's current video model, and the node is a near-clone of its base sibling - which is exactly the point. Same inputs, same VIDEO output, same Runware cloud plumbing, different (bigger, pricier) model on the far end. If you've read the base Wan3.0 page, you already know how this works; what you're really deciding is whether Alibaba's premium tier earns its per-second rate on your material.
Given how the Wan line got here, a little skepticism is fair. Alibaba closed Wan's weights at 2.5 and the community's verdict on those closed releases was lukewarm-to-hostile - the memorable reaction to Wan 2.7 was a thread titled "Its shit tbh" whose top comments said 2.2 looked better. Prime tiers are the vendor's answer to that skepticism: pay more, get the good one. Whether Wan3.0 Prime clears that bar is something you test, not something you take on faith, and this node makes testing cheap because the workflow is identical to the base - swap the node, reuse the prompt and seed, compare.
The full input list carries over unchanged: positivePrompt (sentence-style, with Image 1/Video 1 asset references supported), referenceImages and frameImages for image-to-video and frame conditioning, referenceVideos for video-to-video, settings.audio on by default for native synchronized sound, and settings.promptExtend defaulting on with its reproducibility caveat. size is the preset dropdown (480p to 1080p, multiple aspects, or from-input), plus seed, numberResults, outputFormat (MP4/WEBM/MOV), outputQuality, and the safety group with its fast/full moderation modes. Duration lives in advanced_json, per the tooltip. Output is a native VIDEO.
Install
Every Runware node shares the one-time pack setup: search Runware in ComfyUI Manager and install, or clone https://github.com/Runware/ComfyUI-Runware into ComfyUI/custom_nodes and pip install -r requirements.txt, then restart. Provide an API key from the Runware dashboard - Settings → Runware API key, RUNWARE_API_KEY, or runware auth login - and you're done. There are no model files to download, which is the entire arrangement: Wan3.0 Prime has no open weights, and Runware is the only way most people will ever touch it from ComfyUI.
The honest cost read
This is a closed frontier model rented per call, so treat it like one. Test on a couple of cheap short clips first, watch the per-run cost on the node's title bar, and only commit to Prime when the base tier demonstrably comes up short on your subject. Your prompt, reference images, and videos all leave your machine and pass through Runware's moderation - with no local Wan3.0 to fall back on, that's the price of using it at all. If "newest Alibaba video" isn't specifically what you need, an open-weight model like LTX-2.3 runs locally for the cost of electricity and never sends your footage anywhere.
Inputs (17)
| Name | Type | Default | Description |
|---|---|---|---|
| referenceImagesopt | IMAGE | — | |
| frameImagesopt | IMAGE | — | |
| referenceVideosopt | RUNWARE_REFERENCEVIDEOS | — | |
| positivePromptopt | STRING | Text prompt describing elements to include in the generated output. Reference assets in array order as `Image 1`, `Video 1`, and `Audio 1`, numbered from one. | |
| sizeopt | COMBO | 1080p (16:9) | Output size. Only one of resolution or dimensions is sent. |
| seedopt | INT | 00–2147483647 | Random seed for reproducible generation. When not provided, a random seed is generated in the unsigned 32-bit range. |
| numberResultsopt | INT | 11–4 | Number of results to generate. Each result uses a different seed, producing variations of the same parameters. |
| settings.audioopt | BOOLEAN | true | Generate native audio aligned with visual content. |
| settings.promptExtendopt | BOOLEAN | true | Enable LLM-based prompt rewriting to expand and clarify inputs. Affects reproducibility. |
| safetyopt | BOOLEAN | false | Enable to set safety. Off uses the model's default. |
| safety.checkContentopt | BOOLEAN | false | Enable or disable content safety checking. Increases total generation time. |
| safety.modeopt | COMBO | fast | Safety checking mode for video generation. |
| ttlopt | BOOLEAN | false | Enable to set ttl. Off uses the model's default. |
| ttl_valueopt | INT | 60 | Time-to-live (TTL) in seconds for generated content. Only applies when `outputType` is `URL`. |
| outputFormatopt | COMBO | MP4 | File format for the generated video. |
| outputQualityopt | INT | 9520–99 | Compression quality of the output. Higher values preserve quality but increase file size. |
| advanced_jsonopt | STRING | Optional JSON merged into the request. For: inputs.referenceAudios, inputs.documents, inputs.urls, duration |
Outputs (1)
| Name | Type | Description |
|---|---|---|
| video | VIDEO | — |