Latest models
- Seedance 2.0 (February 12, 2026) Current
A multimodal audio-video model that takes text, image, audio and video inputs and generates clips of 4 to 15 seconds with sound. Source: ByteDance Seed ↗
Checked October 7, 2026. See the full AI model tracker.
What we like
- Takes text, image, audio and video as inputs together
- Generates audio and video jointly, including stereo sound
- Clips from 4 to 15 seconds
Watch out for
- Access depends on the platform you use
- Check the platform for supported resolutions and prices
- Expect to iterate to get a usable take
Seedance is ByteDance Seed’s video generation family. ByteDance Seed published the official launch of Seedance 2.0 on February 12, 2026.
What is current
Seedance 2.0 is a unified multimodal audio-video model. ByteDance says it accepts four input types (text, image, audio and video) and can combine up to 9 images, 3 video clips and 3 audio clips with natural-language instructions, generating audio and video together at 4 to 15 seconds. It is available globally through ByteDance’s Dreamina platform. See the model tracker.
How prompting works
Name each reference by its role: which image is the character, which clip gives the camera movement, which audio sets the rhythm. Then describe the scene in a sentence or two.
Who should use it
Creators who start from references such as characters, shots and soundtracks rather than only text. Compare it with Veo and Kling.