Supported AI Models
Choosing a Model
The Revise agent can run on a range of leading AI models from multiple providers. Open the model picker at the bottom of the chat box to switch between them - each one has different strengths, speeds, and costs, so it's worth experimenting to find what you like.
The model picker currently offers:
| Model | Provider | Highlights | Plan |
|---|---|---|---|
| Gemini 3 Flash | Fast · Long context | All plans | |
| GPT-5.4 Mini | OpenAI | Low cost · Web search | All plans |
| GPT-5.4 | OpenAI | Well balanced · Web search | All plans |
| GPT-5.5 | OpenAI | Well balanced · Web search | All plans |
| GPT-5.5 Pro | OpenAI | Well balanced · Web search | Revise Pro |
| Grok 4.1 Fast | xAI | Real-time 𝕏 search | All plans |
| Grok 4.1 Reasoning | xAI | Real-time 𝕏 search | All plans |
| Claude 4.5 Haiku | Anthropic | Extremely fast · Web search | All plans |
| Claude 4.6 Sonnet | Anthropic | Web search | All plans |
| Claude 4.6 Opus | Anthropic | Web search · High reasoning | Revise Pro |
Each model in the picker also shows a cost rating - a rough guide to how quickly it uses your AI allowance. Faster, cheaper models let you do more before reaching your limit.
Web Search
Many models can search the web when a task calls for current information. When the agent is searching, you'll see its status update in the chat. No setup is needed - the agent decides when a search will help.