Opening the model picker
In chat: Click the model name displayed in the bottom-left of the prompt bar. The model picker opens as a panel or modal overlay. In Studio (Images, Video, Audio): Click the model dropdown in the Studio controls section. Studio pages have their own model selector scoped to the relevant model type (image, video, or audio models only).Model picker interface
The picker displays:- Search bar at the top — type to filter models by name
- Provider filter — tabs or sidebar listing all 19 providers; click to filter to a single provider’s models
- Model list — each model shows its name, Standard or Premium badge, context window size, and capability icons
- Capability indicators — icons showing whether a model supports vision, tool use, or extended thinking
- Favorites section — models you’ve starred appear pinned at the top for quick access
Selecting a model
Open the model picker
Click the model name in the bottom-left of the prompt bar in any chat. The picker opens.
Browse or search
Scroll through the model list, or type in the search bar to filter by name. For example, type “claude” to see all Claude models, or “gemini flash” to narrow to Flash variants.
Filter by provider (optional)
Click a provider name in the filter sidebar or tabs to show only that provider’s models. Useful when you want to compare all models from a specific provider.
Click a model to select it
Click the model name to select it. The picker closes and the model name updates in the prompt bar.
Pinning favorites
To pin frequently used models to the top of the list:- Hover over any model in the picker
- Click the star icon that appears
- The model moves to the Favorites section at the top of the picker
Understanding model badges
Each model in the picker displays a badge indicating its tier:| Badge | Meaning |
|---|---|
| Standard | No monthly cap — available on all plans |
| Premium | Counted against your monthly quota — requires Pro+ plan |
Context window and capabilities
The model picker shows key information for each model at a glance:- Context window: shown in tokens (e.g., 128k, 200k, 1M) — this is how much text the model can process in a single conversation
- Vision: indicates the model can analyze images attached to your message
- Tools: indicates the model supports tool/function use
- Thinking: indicates the model supports extended reasoning (reasoning models)
Reasoning models and thinking levels
When you select a reasoning model (like o3, o4-mini, DeepSeek Reasoner, or Claude with extended thinking), you may see a thinking level slider appear in the prompt bar. This controls how deeply the model reasons before responding:| Level | Behavior |
|---|---|
| Low | Quick thinking — faster responses, lighter reasoning |
| Medium | Balanced — good for most complex tasks |
| High | Deep thinking — most thorough, slower, uses more quota |
Switching models mid-conversation
You can switch models at any point in an existing conversation:- Click the model name in the prompt bar
- Select a different model from the picker
- Continue the conversation — your next message uses the new model
Switching models in a chat only affects that chat. Other open or saved chats continue using their own selected models. Model selection is per-chat.