No. Standard Wan 3.0 is selected when the page opens, but the model picker remains available. Input controls, supported duration, resolution, audio options, and the credit estimate update for the active selection, so confirm the current setup before submitting.

Begin with text when the visual identity and motion may still be invented. Use a first frame or frame pair when the opening look or both endpoints must stay in control. Use multimodal references when separate images, motion clips, and audio cues need distinct responsibilities inside one brief.

Name the decision that must survive the generation. Start with text when composition and motion may be invented, with one or two frames when an approved visual state must hold, and with multimodal references when identity, movement, and sound need separate sources. That decision is more useful than choosing a route by habit.

Give the first pass one visible job and describe it in time order: opening state, main action, camera behavior, development, ending beat, and sound. Define what must remain stable and what may change. You should be able to watch the result once and say which instruction passed or failed.

Use the frame to lock only the visual decision you need: identity, product geometry, composition, or the opening pose. Describe the motion separately and keep the first test restrained. Add a last frame only when the destination itself matters; otherwise the extra endpoint can make a simple movement harder to judge.

Assign one job to every uploaded asset, then connect it with the matching @ token. An image might define the subject, a clip might demonstrate movement, and an audio file might set a cue. Wan 3.0 accepts up to 10 images, 5 videos, and 5 audio files, but unused or competing references make the result harder to diagnose.

Use a short generation to validate identity, one action, a camera move, or a transition before asking for a longer sequence. Move toward 30 seconds only when the scene needs connected development and the prompt can describe those beats clearly. Standard Wan 3.0 supports 2–30 seconds, with 480P, 720P, and 1080P options.

Keep the creative brief, source assets, and review criterion unchanged if you want a meaningful model comparison. Then confirm the active input controls, duration, resolution, aspect ratio, audio option, and credit estimate because those settings can change with the selected model. Compare one variable at a time.

First watch at normal speed and decide whether the intended scene reads. Then inspect the anchor you chose: identity, frame continuity, reference behavior, camera path, ending state, or audio timing. Record the clearest mismatch and change the instruction, source, or setting most likely to affect it before generating again.

Keep the final prompt, selected model, input mode, duration, resolution, aspect ratio, audio choice, source filenames, and the single change tested in that version. My Creations helps you revisit task status and completed results, while a compact production note makes the successful direction easier to reproduce.