Models and who pays

Pick the model and reasoning effort in the composer, see who pays for each turn, and read your activity on the Usage page.

Every reply comes from a model you pick in the composer. Some models come with June and June pays for them; others run on your own ChatGPT or Claude plan, or on an API key.

The model picker and the reasoning effort sit at the bottom of the composer.

Pick a model

Open the model picker in the composer, or press ⌘⇧M (Ctrl+Shift+M). Models are grouped by who runs them:

Group	Shows when	Who pays
June Models	Unless a funding source pays	June
ChatGPT	You connected a ChatGPT plan and no funding source pays	Your ChatGPT plan
Claude	You connected a Claude plan and no funding source pays	Your Claude plan
A source's name	Your space, or the organization, has a funding source	That source: an API key or a shared plan

When a funding source pays in a space, its group is the only one there: June Models and your own plans are not offered.

June remembers your pick across chats and devices. In an organization a funding source pays for, a pick holds for the chat you made it in and leaves the pick for your own chats as it was. Once a ChatGPT or Claude plan has answered in a chat, that chat stays with that plan: the other groups are shown but cannot be picked, and you start a new chat to switch.

Reasoning effort

Next to the model, pick how long it thinks before it answers: Low (the default), Medium, High or Extra high. Higher effort is slower and suits harder questions. Not every model offers every level: Extra high shows only on the models that support it. June remembers your choice.

Who pays for a turn

A funding source picked in the space pays for every turn there. Otherwise the model's group decides: June pays for June Models, and your plan pays for its own models. When no plan of yours can answer, June answers on its fallback model and pays, and the reply ends with "Using GPT-5.6 Luna. Attach a ChatGPT or Claude subscription to use your plan."

A schedule is paid the same way in the space it runs in: by the space's funding source if it has one, otherwise by the person who set it up, as for a reply of their own.

Your activity on the Usage page

The Usage page reports everything June ran for you: model tokens, time in June's sandbox, storage and embeddings, by day. Narrow it to the last 7, 30 or 90 days or this month (the default), and filter by funding, provider, model, channel or operation.

It is an activity report, not an invoice: nothing on it is billed and no limit comes from it.

Your plan's limits

Your ChatGPT or Claude plan sets its own usage windows, and June works inside them. While one of the plan's models is picked, the meter next to the pickers shows how much of each window is used and when it resets; Use your own plan has the details.

What June does not do

It does not sell credits or show a price list: there is no pricing page.
It does not mix sources in one turn.
It does not make images unless the reply runs on a ChatGPT plan.