Choose an LLM Provider and Model
Testronaut uses an LLM to reason about the interface, choose browser actions, and determine whether a mission succeeded. Choose a provider during testronaut --init, then override the model when a particular mission needs a different balance of capability, speed, or cost.
Need credentials first? Use the official provider links in Set up model API keys.
A practical starting point
If you are new to Testronaut:
- Choose a provider for which you already have API access.
- Start with that provider's recommended general-purpose tool-using model.
- Run one representative mission.
- Review its result, turn count, latency, and token use.
- Move to a faster model for simple smoke tests or a stronger model for difficult journeys.
Model availability, names, prices, and rate limits change independently of Testronaut. Treat the provider guides as selection guidance and confirm current pricing with the provider.
Provider guides
Where model selection fits
Choose and validate a model before investing in advanced mission patterns such as chaining missions or increasing turn limits. This gives you a known execution baseline and makes cost or reliability changes easier to interpret.