What’s new
- Billed as the fastest way to use Runway’s most powerful video model: Runway’s launch post described it as "the fastest way to generate with our most powerful video model yet," with scale and reliability kept at the level of the previous Gen-3 Alpha Turbo.
- A 10-second clip in about 30 seconds: Runway’s own launch post put it plainly — generating a 10-second video now takes just 30 seconds, faster than standard Gen-4, which Runway pitched for rapid iteration and creative exploration.
- More flexible duration via the API: since October 2025, the Runway API lets you pick any duration from 2–10 seconds with Gen-4 Turbo, instead of being limited to the product’s two fixed options (5s / 10s).
What it’s good for
- Animating an image whose composition and style are already locked in — a product shot, a character portrait, a keyframe.
- Quickly testing camera moves or directions of motion before committing to a fully detailed final shot.
- Iterating on the same source image with different motion prompts until one feels right.
Prompting tips
- Prompt for motion only, not the picture: the image already sets composition, lighting and style, so Runway says the text prompt should be “almost entirely focused on describing the desired motion.”
- Use positive phrasing: say what you want rather than what you don’t — Runway’s prompting guide notes that negative phrasing can produce unpredictable results.
- Structure it as “camera move + subject action”: for example, “the camera slowly pushes in as the subject turns.”
- State it explicitly if you want a static camera: add a line like “the camera remains perfectly still.”
- Add “continuous, seamless shot” if you’re getting too many cuts: this cuts down on unwanted scene breaks.
- Start simple, then add detail: keep your first prompt short, check the result, and iterate from there.
The camera slowly pushes in as the skateboarder takes a running start, launches
into the air, completes a spin, and lands cleanly. Wind lifts the hem of his jacket.
Continuous, seamless shot, no cuts, camera holds steady throughout.
Known limits
- Image-to-video only — no pure text-to-video: Runway Dev’s official model catalog lists only “Image to Video” as a supported task for Gen-4 Turbo, and the Help Center confirms Gen-4-family generations require an input image. For text-to-video, Runway’s current flagship, Gen-4.5, supports it.
- Only two duration presets in the product: 5 or 10 seconds; picking any duration from 2–10 seconds requires going through the API (available since October 2025).
- Text prompts are capped at 1,000 characters.
Using it in Nomi
On the generation canvas, add a video card, pick “Runway Gen-4 Turbo” in the model picker and choose a mode (First frame to video). Add your references, write the prompt and generate. In the storyboard you can pick it per shot. The first time, connect any one of the providers above under Settings → Models.
