OpenArt adds Seedance 2.0 with 9 image refs, 3 videos, 3 audio files

• MayorKingAI's OpenArt launch thread says Seedance 2.0 is live now for Teams and Enterprise customers, while the official OpenArt model page describes the release as a multimodal video generator with native audio and cinematic camera controls.
• According to MayorKingAI's feature rundown, OpenArt is exposing unusually high reference limits for a consumer-facing video tool: up to 9 images, 3 videos, and 3 audio files.
• Magiermogul's early-access reel focuses on low-key dialogue scenes and claims Seedance 2.0 can keep 15 seconds of video, audio, and dialogue consistent, which lines up with Dreamina's official Seedance page pitching native voice, singing, and multi-camera continuity.
• Anima Labs' Arcads post says Arcads added Seedance 2.0 alongside a new Nodals production system and avatar tools, suggesting the model is already being slotted into broader ad and content pipelines.

You can jump straight to OpenArt's Seedance page, browse Dreamina's official tool page, and see the rollout surface in the wild through OpenArt's launch thread, Magiermogul's dialogue-heavy test reel, and Arcads' platform update. The weirdly useful detail is that each wrapper emphasizes a different part of the stack: OpenArt sells control knobs, Dreamina sells an end-to-end studio, and Arcads sells insertion into ad production.

OpenArt's version of Seedance 2.0 is all about input bandwidth. The official page says creators can combine text, images, video, and audio, then upload up to 9 images, 3 audio files, and 3 videos with 15 seconds of total video reference.

That same page also promises synchronized audio generation, lip-sync, and camera moves including tracking, dolly, POV, and rack focus. It is a lot of structure for a single model page, Christmas come early for prompt-control obsessives.

The bigger shift is that Seedance 2.0 is being framed as a scene builder, not a one-shot clip machine. The official OpenArt page says users can extend, merge, or edit clips, while Dreamina's official page leans on "seamless multi-camera storytelling" and continuity between plot and shots.

Magiermogul's early-access reel is useful here because it is not a loud VFX demo. His thread says he specifically tested realistic, character-centered scenes, then claimed the model could keep 15 seconds of video, audio, and dialogue consistent. That is a more interesting stress test than another explosion montage.

MayorKingAI's OpenArt thread is basically a genre spread. Instead of one hero clip, it runs Seedance 2.0 through:

1. a lizard-versus-cat chase
2. a market chase
3. an American football movie scene
4. a Porsche ad
5. an anime fight
6. a transformer fight
7. a dragon attack
8. a sunflower walk

The list matters because it shows where Seedance is being positioned. Not just short films, also ad creative, stylized action, and polished brand-video lookups.

Dreamina's own packaging goes further than model access. The official Dreamina page says creators can use Seedream 5.0 for storyboard images, an AI Agent to plan shots and scenes, and AI Avatar to add digital humans, all inside one workflow.

Dreamina also pitches native voice and singing, style transfer, and locked character consistency across recurring scenes. The page even spells out output settings, including 720p to 1080p resolution and 5 to 12 second durations, which makes it feel more like a production template than a pure model showcase.

Arcads is the clearest sign that Seedance 2.0 is escaping its home turf. Anima Labs' rollout post says Seedance 2 launched there alongside other video models, image models, a new Nodals system for start-to-finish production, and avatar creation for AI influencers with their own voices.

The broader Arcads product site already centers AI ad generation, custom avatars, and a library of more than 1,000 AI actors. Adding Seedance 2.0 there gives the model a very specific job: not just cinematic experimentation, but fast ad assembly inside an existing marketing stack.

Availability is messy, and that is new information in itself. MayorKingAI's launch post says OpenArt has Seedance 2.0 live for Teams and Enterprise users, while the follow-up post sends US and individual users to a waitlist.

At the same time, Magiermogul's Dreamina post says Dreamina has the model on its web app and mobile app, but only as a rollout across select countries and regions. Arcads, by contrast, presented Seedance 2 as available that day inside its platform. Same model family, three very different doors in.