Browse Core Concepts
Every request in Agents is answered by an AI model, and you choose which one. The model menu sits in the composer beside the + button and shows the model in use with its thinking level, such as "Haiku 5.5 Medium." The default handles most work well. Switch when a job needs more depth, or when you want a quicker answer.
Open the model menu
Click the model name in the composer. The menu shows:
- The current model, with a one-line description of what it is best at.
- Thinking, which sets how hard the model works before it answers.
- More models, the full list from Anthropic, OpenAI, Google and xAI.
Pick a model and it answers from your next message on. MorningAI remembers your choice until you change it.
Pick a model for the job
The list grows as new models come out, so think in types rather than names:
| For | Choose |
|---|---|
| Quick questions, short copy and everyday edits | A fast model. Its description says so, for example "Fastest Claude model." |
| Long documents, analysis across several sources, plans with many steps | The newest flagship model from any provider |
| A second opinion on an important draft | A model from a different provider, then compare the two |
Earlier versions stay available when a new model arrives, so a habit or a Playbook built on one keeps working.
What Is in the List
The list changes as providers release new models. As of November 2026 it includes:
| Provider | Models |
|---|---|
| Anthropic (Claude) | Fable 5.2, Fable 5.1, Opus 5.5, Opus 5, Sonnet 5.5, Sonnet 5, Haiku 5.5 |
| OpenAI (GPT) | 6.1 Sol, 6.1 Terra, 6 Astra, 6 Sol, 6 Luna, 5.6 Sol, 5.6 Terra |
| Google (Gemini) | 4 Argon, 3.8 Flash, 3.1 Pro, 2.5 Pro, 2.5 Flash |
| xAI (Grok) | 4.7, 4.6, 4.20 Fast |
Fable 5.2 is the newest addition and sits at the top of the Claude models.
Set the thinking level
Thinking runs from Low to Max. More thinking gives more thorough answers, but replies take longer. Not every model offers every level, so the menu shows the levels the selected model supports.
| Level | Use it for |
|---|---|
| Low | Quick answers where speed matters more than depth |
| Medium | Most work. This is the recommended setting |
| High | Plans, analysis and documents with several moving parts |
| Max | The hardest problems, when you can wait for the answer |
Models in scheduled Playbooks
When you schedule a Playbook, you can pick the model for that schedule, so a weekly report can run on a different model from your everyday chats. See How to Create and Use Playbooks.
Frequently Asked Questions
Does the model change what the Agent knows about my brand?
No. Every model works from the same DNA, Files and connectors, and changes still come to you on approval cards. What differs is how the model writes and reasons.
Can I change models partway through a conversation?
Yes. The model menu stays in the composer, so you can switch before any message.
Does the model menu apply to Studio?
No. The model menu is part of the Agent composer. Studio has its own quality settings, such as Ultra and the QA Agent.
Next Steps
Ask the same question on Medium and then on High, and compare the two replies. That is the quickest way to learn when the extra thinking is worth the wait.
