
Fast reference-driven generation. Same consistency, 2.5x faster.
The speed variant of Runway's image generation pipeline. Turbo generates approximately 2.5 times faster than the standard model at a lower cost, while keeping the reference-driven consistency features. Requires at least one reference image. Use it to iterate quickly on compositions, test reference combinations, and explore directions before committing to a final Gen-4 render.

“Same headphone, now on a dark wood desk next to a coffee cup, warm morning light, lifestyle product photography”

“Same character, now in a winter coat walking through a snowy Stockholm street, blue hour, editorial fashion”

“Same house, now at night with interior lights glowing warm, moonlit ocean, long exposure feel”

“Same captain character, now younger, clean-shaven, standing on a sunny dock, bright and hopeful mood”

“Same outfit flat lay, now on a dark slate surface with dramatic overhead lighting, premium menswear”

“Same ramen shop, now daytime, quiet afternoon, sunlight cutting through steam, peaceful mood”
Type a detailed prompt or upload a reference sketch, photo, or mood board.
Pick your resolution and aspect ratio. See the credit cost before you generate.
Your image is delivered in seconds. Download, iterate, or pipe into video.
Jump into the Studio and start generating. Plans from £10/month.
Choose a PlanRunway Image Turbo is built for the exploration phase. It requires at least one reference image and generates approximately 2.5 times faster than the standard Gen-4 model. The reference system works identically: character, style, and location references are all supported. The speed gain comes from optimised inference, not a reduced model. Visual quality is strong for concept work, with the difference from the standard model showing only in fine texture detail on close inspection.
Turbo makes high-volume iteration affordable. Test a product reference across twelve background environments. Try a character in eight different costumes. Explore a location reference under morning, midday, golden hour, and blue hour lighting. Each test returns in seconds. Once you find the direction, switch to Gen-4 Image for the final renders at full quality.
The requirement for a reference image is not a limitation. It is the point. Turbo is designed for workflows where you already have a visual anchor: a product photo, a character sketch, a mood board image, a brand style frame. You are not generating from nothing. You are generating variations of something you already have. That constraint is what makes the output consistent and the iterations meaningful.
Turbo supports the same reference system as Gen-4 Image: character, style, and location references up to 3 images. The speed advantage lets you test more combinations in less time, finding the right reference mix before committing to final renders.
Every Turbo generation starts with a reference image. This is by design. The model excels at producing variations of an existing visual, not generating from scratch. Feed it a product photo and describe new contexts. Feed it a character and describe new scenes. The reference anchors consistency across every output.
A small indie studio building creative tools the way they should be built. No VC theatre, no funnel games, no faceless support.
Professional image generation. Plans from £10/month.