Skip to content

Model Combos

Combos group several models into a single named entry you can select anywhere the app asks for a model — chat, AgentUI or Batuta. If one model fails, hits a rate limit or runs out of quota, the combo fails over to the next one automatically.

  • Resilience — a provider outage or quota exhaustion never blocks work; the next model in the combo takes over.
  • Cost control — order models from cheapest to most expensive and let cheaper models handle what they can.
  • One choice, everywhere — configure the combo once and reuse it across chat, agents and orchestrations.

Don’t know which models to combine? Describe your need in natural language — “a cheap combo for code review with a strong fallback” — and the assistant recommends and orchestrates a model hierarchy with failover based on your connected providers.

The app monitors quota per account and shows it where you choose models:

  • Plan tags — Pro / Free indicators next to each account.
  • Real-time percentage — current usage at a glance.
  • Detailed modal — per-model quota and renewal time.

Accounts approaching exhaustion (from 95% usage) are paused automatically, and the next connected account assumes — protecting you from mid-task failures.

The model selector only shows providers that actually have active credentials or configured combos — no dead entries, no guessing.