Seedance 2.0

Overview

Seedance 2.0 is ByteDance's video generation model, released February 10, 2026, and it has held the #1 spot on the Artificial Analysis Video Arena for both text-to-video and image-to-video since launch, ahead of Google Veo 3.1 and Kling 3.0 — at a significantly lower cost. It's the same model powering Dreamina, Jimeng, Doubao and CapCut under different brands. Its real differentiator is multimodal input: up to 9 images, 3 videos and 3 audio files as references in a single generation, addressable directly in the prompt (@Image1, @Video1...). You can replicate choreography from a reference video, lock a character or product across the whole clip, and get natively synchronized audio with no post-production. Clips run 4 to 15 seconds, up to 1080p (a 4K upgrade was announced in June 2026), with multi-shot editing in a single pass. Access: global API via fal.ai since April 9, 2026 ($0.30/s at 720p standard, $0.24/s fast tier), consumer interface via Dreamina (~$18/month, free tier of 225 tokens/day). Two caveats: the model is not available in the United States, and it faces cease-and-desist letters from Disney and Paramount over IP infringement — a legal risk to factor in for commercial use. Its successor Seedance 2.5 (30-second clips, 50 references) was announced June 23 and lands in July 2026.

Skill profile

Who is it for

Ideal use cases

• Ads and short-form social content
• Cinematic pre-visualization (animated storyboards)
• Product videos with brand consistency via references

Access & availability

Key specifications

Privacy