Unlimited generations queue, because unlimited bounds what you spend and not how fast it arrives. Two things decide when your output lands.
The two delays
What your plan sets
Unlimited generations run 2x faster on Basic and 4x faster on Pro, Pro Plus, and Max. Pro Plus and Max can also queue up to 10 generations at once. What the plan does not set is how long the model itself takes, so a high-quality video model still takes minutes on every tier.
Generation time is set by the model
The spread across models is wide. A light image model finishes in seconds; a top-end video model takes minutes before a queue is involved at all. If you are still iterating, iterate on something cheap and move up for the final render.
Three ways to get output sooner
- Fast Mode. Spends Gems to run one job at the front of the queue, on any model. See Fast Mode.
- A lighter model. The gap between the heaviest video models and the lighter ones is large, in time and in Gems.
- A lower resolution or a shorter clip. Both reduce the work, and both keep you inside your plan’s unlimited ceiling.
When it is actually a problem
A job that sits without starting for far longer than usual, on a model you run regularly, is worth reporting instead of waiting out. Send the model, the settings, and the time to mage@mage.space. Last modified on September 14, 2026
