Which models can I use?

The model menu lists what you have connected: your Claude plan, your ChatGPT plan, and your API keys. With more than one connected, you see each group. What a model costs is between you and the company that runs it; Flockbay charges nothing for it.

On your Claude plan, the Claude tab lists the models Claude Code offers your account, under their real names. Signed in with ChatGPT, the menu lists that account’s models under ChatGPT. With an API key, it lists the models that key reaches: that company’s own models on an OpenAI or Anthropic key. None of them are locked.

The list changes as new models come out, so it is not written down here; the menu is always the current list.

Some models let you choose how long they think before they act. A Claude model shows an Effort control under the list, with the levels Claude Code gives for that model. A ChatGPT or API-key model that supports it shows Reasoning, with Low, Medium and High. Higher settings help with harder changes and take longer. Models without the control do not show it.

Not every model can look at a picture. When the chat’s model cannot, the attach button offers only text files, a pasted picture is refused with a sentence that says why, and a picture already in your message holds Send until you remove it or pick a model that can see.

If you connect your own agent over MCP, it runs on whatever model you set up in that agent, and this menu does not apply to it.

With nothing connected, a new account has Demo: a small, low-quality model on a small free allowance, for trying the app.

Flockbay also has optional plans that run the AI for you. A plan opens a band of models chosen by what they cost to run, and its first row in the menu is Auto. The most expensive models, like GPT-5.5 Pro, are on no plan and run only on your own key.