11 models,
5 providers
One picker, every major model. Guests can try the fast ones; holders get the full list.
| Model | Provider | Cost | Who can use it | Good for |
|---|---|---|---|---|
| GPT-6 Luna | OpenAI | Cheap | Guests and holders | Fast everyday answers. |
| GPT-6.1 Terra | OpenAI | Standard | Holders | Strong all-rounder. |
| GPT-6 Astra | OpenAI | Premium | Holders | Hardest problems, longest thinking. |
| Claude Haiku 5.5 | Anthropic | Cheap | Guests and holders | Quick, careful replies. |
| Claude Sonnet 5.5 | Anthropic | Standard | Holders | Writing and code. |
| Claude Opus 5.5 | Anthropic | Premium | Holders | Deepest analysis. |
| Gemini 3.5 Flash Lite | Cheap | Guests and holders | Fast, with long context. | |
| Gemini 4 Argon | Standard | Holders | Long documents and reasoning. | |
| Grok 4.7 | xAI | Standard | Holders | Direct, current-events savvy. |
| DeepSeek V4 Flash | DeepSeek | Cheap | Guests and holders | Lowest cost, capable. |
| DeepSeek V4 Pro | DeepSeek | Standard | Holders | Step-by-step reasoning. |
Costis a rough guide: a cheap reply usually costs a fraction of a cent, a premium one a few cents or more, depending on length. You're charged what each reply actually costs. Guests get 10 free messages on cheap models. The list changes as providers release new models.