Primary navigation

Speed

Increase speed without sacrificing intelligence

ZeroTwo has no separate priority speed tier. You get faster responses by lowering the reasoning level or by choosing a lighter model.

Fast mode

In the desktop app, Toggle Fast mode is an action in Settings > Shortcuts, and it ships without a key assigned. When you trigger it, the composer's reasoning control jumps to the lowest level that the current model and mode offer. Triggering it again returns to the highest level. It only changes the reasoning level: the model stays the same and no priority processing is involved.

Reasoning level

Each model has its own set of reasoning levels in the composer dropdown, and the lower levels respond sooner and spend fewer tokens. In Work and Code the dropdown shows the model's native ladder, for example Low, Medium, High, X-High and Max for GPT 6.1 Sol. Plain Chat shows None, Standard and Extended. Reasoning can also be changed with /reasoning in the slash menu.

Model choice

The model switcher lists credit-cost chips next to each model, so you can pick a cheaper, lighter model for quick work. Auto picks a model for you; it routes to GPT 6 Luna for chat and agent modes. See Models for the full list.