ByteDance, the creators of TikTok, just dropped the best video gen model in the world.

It’s #1 on the Artificial Analysis leaderboard, 50-100 ELO points over Veo 3 for both text-to-video and image-to-video.

Can do accurate physics, consistent characters and complex scenes.

It’s called Seedance.

Use it here: https://lnkd.in/gzJ2Z48z

Kling 2.1 and Veo 3 do native audio, not the rest. Wild to me that each of the top 4 came out last MONTH.

Prompt: "The team swings couch pillows at each other in gentle slow motion. A final soft whack lands; the struck founder stops mid-swing and quietly turns to smooth gray stone while feathers drift past"

With the classic OpenAI launch image.


This post was originally shared by on Linkedin.