Runway launches Seedance 2.0

Runway just turned one of the internet’s hottest video models into a button inside its own app, so a paying user can now drop in a prompt, an image, a clip, or audio and get back a short video with sound in the same workflow. That matters because most artificial intelligence video tools still make you stitch pieces together by hand, while Seedance 2.0 is pitched as one model that can handle multiple shots, synchronized sound, dialogue, and character consistency in a single generation. Runway says Seedance 2.0 is a third-party model, not one of Runway’s in-house Gen-4 systems, so this launch is also a marketplace move: keep creators inside Runway even when the model they want came from ByteDance. The basic pitch is simple: instead of filming 10 separate clips and editing them into a sequence, you can ask for a 5-to-15-second scene with several camera cuts already baked in. Runway lists support for text, image, video, and audio inputs, with outputs at 480p or 720p and aspect ratios from widescreen 21:9 to vertical 9:16. Runway is putting that tool behind its paid wall, with help documentation saying Seedance 2.0 is available on Standard plans or higher. Its pricing page lists Standard at $12 a month billed annually and Pro at $28 a month billed annually, with higher tiers above that. To speed adoption, Runway is also dangling a time-limited discount: 50% off three months with the code SEEDANCE, ending April 13 at 9 a.m. Pacific Time. That tells you this is not a quiet feature drop; it is a push to get people testing the model immediately. Runway’s sales angle is advertising as much as filmmaking. On the Seedance page, it tells marketers they can start with product images, add a script, and generate multiple versions for different channels without hiring a freelancer or agency for each cut. That lines up with Runway’s Big Ad Contest, which asked paid users to make a 30-to-60-second spec commercial for one of seven fictional products and offered up to $100,000 in cash prizes. The contest ran from March 18, 2026 to April 2, 2026 under United States law, with entries required to be made inside Runway. The company is also leaning on creator proof instead of benchmark charts. Higgsfield, another platform offering Seedance 2.0, advertises the same model for short films, music videos, ads, and social clips, with claims of multi-camera storytelling, frame-level precision, and native audio sync. There is one visible constraint in Runway’s own help page: Seedance 2.0 is likely to moderate inputs that depict realistic humans, and Runway recommends stylized characters or avoiding photorealistic faces in some image-based workflows. So even as the model gets closer to “type it and screen it,” identity and likeness controls are still shaping what creators can actually make. The bigger picture is that Runway now sells access to its own video models, outside models like Seedance 2.0, editing tools, audio tools, and an application programming interface in one place. The fight is no longer just over who has the best model; it is over who becomes the default studio where creators never need to leave.