Choosing a Model

By Task

Task	Use	Where
Text	A chat model	Model Families
Step-by-step problems	A reasoning-capable model	Browser App: the effort control. Plugin: Reasoning. API: Reasoning
Images	A vision-capable model	Browser App, Plugin, API
Documents	Any model, with the file attached	Browser App, Plugin, API
Current information	Any model, with web search on	Browser App, Plugin, API
Several answers at once	Model Fusion	Browser App, Plugin

Smaller models are faster and usually enough. Reasoning costs extra tokens and time.

By Privacy Level

The level comes with the model, so this is the same choice. Take the lightest level that fits the work.

Level	Take it when	What you get
Anonymous	You want a frontier model for ordinary work	Frontier models, no identity attached
Private	You want no data retained	Zero data retention upstream
Confidential	You want proof of where the call ran	Sealed hardware with attestation you can verify
Encrypted	The prompt must stay unreadable in transit	Confidential plus encryption on your device. Browser App and API today; Plugin coming soon

A call that cannot run at the level you asked for fails rather than dropping to a weaker one.

Nito Auto

In the Browser App, Nito Auto picks a model for you and is included in every paid plan. In the Plugin, a built-in default answers until you set one. See Choosing a Model in the Browser App and in the Plugin.

Model List

The catalog by privacy level.

Model ID Convention

How a name shows its level.

Plans

Which models each plan can use.

​By Task

​By Privacy Level

​Nito Auto

​Related