Choose a model
How to pick which model runs a Copilot chat, what each one is suited to, and where to set the one new chats start with.
Copilot runs on a language model, and you can pick which one. It is the model that reads your request, plans the work, and talks back to you. Which image or video model makes the media is a separate setting, under 'Model preferences' in the prompt bar.
The models
| Model | Suited to |
|---|---|
| Auto | Picks the best model for each task. The default |
| Sonnet 5.5 | Strong all-rounder for most tasks |
| Haiku 5.5 | Fast, affordable for quick tasks |
| Grok 4.7 | Fast, strong on long tasks |
| GPT 6.1 Sol | Efficient reasoning for complex tasks |
| GPT 6 Astra | Top reasoning for the hardest tasks |
| GPT 6.2 Astra | Sharper Astra for the hardest tasks |
| GPT 6.1 Terra | Balanced, sharp on visuals and writing |
| GPT 5.6 Terra | Balanced for everyday work |
| Fable 5.2 | Creative stories and ideas |
| Opus 5.5 | Strong reasoning for complex work |
| Gemini 4 Argon | Long memory for large projects |
| Gemini 4 Neon | Capable Gemini for everyday work |
| Kimi K3.1 Pro | Strong agent at a lower price |
Leave it on 'Auto' unless you have a reason not to.
- GPT 5.6 Terra, Fable 5.2, Opus 5.5, Gemini 4 Argon and Gemini 4 Neon need a paid plan. On the Free plan they show a lock and the note 'Available on paid plans'.
- A model an admin has turned off for your organization shows 'Not available in this workspace'.
- Each model shows two ratings as dots: 'Capability' and 'Cost'.
Credits
Each reply uses credits, and the model sets the rate. Auto is usually the cheapest choice. A model you pick is always charged, which the dropdown notes under the list: 'Switching from Auto to any specific model will cost you the relevant credits'. Auto can be charged too, depending on your account.
Changing the model partway through a chat asks you to confirm first: the new model reads the whole chat before it answers, so its first reply costs more. Pick the model when you start a chat to avoid it.
Typical credits per reply for every model are under Copilot reply costs.
Effort
Below the model list, 'Effort' sets how long the model thinks before it answers: 'Low', 'Medium', 'High', 'Extra high' or 'Max'. 'Default' uses the model's own level. Higher effort thinks longer and uses more credits. If a model does not offer the level you picked, the chat runs at its default.
Pick one for a chat
- Open Copilot
- Click the button at the bottom right of the prompt bar. It reads mode first and model second, so 'Auto · Sonnet 5.5' is automatic generation running on Sonnet 5.5
- Under 'Model', open the dropdown and pick one. A checkmark marks the active model

The choice belongs to the chat, so chats can run on different models at once, and it holds until you change it.
Set the default for new chats
To make the current model the default, hover a model in the menu and click 'Make Default'. Or:
- Open 'Preferences'
- Beside 'Default chat mode', open the dropdown
- Pick the model under 'Model', and the mode under 'Mode' while you are there
New chats start on that model. Changing one chat's model does not change the default.
That's how you choose a model on Morphic. Most work is fine on 'Auto'; pick one yourself when a chat calls for speed or for more reasoning.
If you have any trouble or have any questions, feel free to reach out. We're happy to help. Write to us at support@morphic.com.