Selecting a model for a conversation
The chat input toolbar includes a model picker (the sparkle icon with the model name) that lets you choose which AI model powers the current conversation. The available models as of this writing are:
Claude Sonnet 4.6 (Anthropic), fast and balanced. The default for most workspaces.
Claude Opus 4.8 (Anthropic), Anthropic's highest-capability model; includes extended thinking. Indicated by a sparkle icon in the picker.
GPT-5.5 (OpenAI)
GPT-5.4 mini (OpenAI)
Gemini 2.5 Flash (Google)
Models that support extended thinking are marked with a sparkle icon next to their name in the dropdown. The model is fixed for the life of a conversation; to switch models start a new chat.
Choosing an assistant
Next to the model picker is an assistant picker (the person icon). Assistants give the model a persona and set of instructions that shape how it responds. The built-in assistants are:
General, default, all-purpose responses
Researcher, structured analysis and synthesis
Marketer, copy and content-oriented responses
Sales, deal and prospect-focused responses
Recruiter, hiring and candidate-oriented responses
You can also create custom assistants with your own instructions. Click Create new assistant at the top of the assistant picker dropdown, or manage your assistants at Profile under the AI Assistant section. Custom assistants appear under Your Assistants in the picker.
Setting a workspace default model
To avoid picking a model every time, set a default at Profile > AI Assistant. Under the Default AI model section, choose the model you want from the dropdown and the setting saves immediately. Individual chats can still override the default using the model picker in the input toolbar.
Related articles
The fastest way to reach us is the live chat in the app (the chat bubble in the bottom corner). You can also email [email protected].