Models
Video models behind AdaptLane
Which model actually renders each shot
AdaptLane routes each job to a model based on resolution and length: shorter shots go to the cheaper model, long shots and 2K go to the more capable one. Here is every model currently wired in, and what it is responsible for.
ByteDance
Seedance 1.5 Pro
The default workhorse: everyday shots at 480p to 1080p
- Handles:
- 480p / 720p / 1080p, up to 12s
ByteDance
Seedance 2.0
Longer shots: picks up where 1.5 Pro runs out of length
- Handles:
- Around 15s
ByteDance
Seedance 2.5
Extra-long shots: only model that handles 20-30 seconds
- Handles:
- 20-30s
MiniMax
MiniMax Hailuo 3 (2K Cinematic)
2K Cinematic exclusive: native 2560x1440 output
- Handles:
- 2K Cinematic only
How the platform picks a model
You choose resolution and length; the platform filters by capability and calls the matching model. The same prompt at 30 seconds or at 2K may be rendered by a different model β that is how availability and cost get balanced.
Start generating