CodeMuxHow do adaptive model levels work?
New work starts at your configured default. The model can continue at the next level or Maximum without repeating completed tool calls or losing the task's Undo history.
Models & providers
GPT, Claude, DeepSeek, Grok, and Gemini in one coding workflow.
CodeMux ships with a managed catalog from OpenAI, Anthropic, DeepSeek, xAI, and Google. Configure Fast, Strong, and Maximum once, choose a starting level, and let the current task move upward when needed.
Catalog snapshot reviewed . The catalog inside the extension is always the current one.
Managed model catalog
| Provider | Models |
|---|---|
| OpenAI | GPT-6 Luna, GPT-6.1 Sol, GPT-6 Astra |
| Anthropic | Claude Opus 5.5, Claude Opus 5, Claude Sonnet 5, Claude Fable 5 |
| DeepSeek | DeepSeek V4.1 Flash 1017, DeepSeek V4 Pro |
| xAI | Grok 4.6, Grok 4.7 |
| Gemini 3.8 Flash, Gemini 4 Argon |
Availability, limits, and compatibility can change. A model being listed does not mean every task or step will use it.
Can I use my own API provider?
Yes, if the endpoint is compatible. Compatibility depends on the endpoint and model, so not every API service is supported. Your provider bills those requests directly. See how managed and API billing differ.
Is CodeMux made by a model provider?
No. CodeMux is an independent product, not affiliated with or endorsed by OpenAI, Anthropic, DeepSeek, xAI, Google, or Microsoft. Read the product facts and privacy notice.