Maestro
Maestro gets smarter
Every task Maestro routes teaches it something new: which plan worked, which model won, where it ran fastest, what it cost, and how many tries it took.
More tasks, better routes. Better routes, better price, better speed, fewer tries. Better price and speed, more tasks.
What it learns from each task
- Which plan worked: how the job was broken into tasks.
- Which model won: did the result pass its check?
- Where it ran fastest: on which cloud, at what time.
- What it cost: at that moment's prices.
- How many tries it took: and which model got it right.
New models, used the moment they are best
New models arrive every week. Maestro learns what each one is good at, and uses it the moment it becomes the best choice. A model joins the board once it proves itself on real tasks.
What this means for you
The stack you use next year is smarter than the one you use today, and you did nothing to make it so. Your AI team gets every improvement on its next task.
Was this page helpful?Thank you. Tell us more at apply@zbstack.xyz.