Choose models and manage AI accounts

Your squad runs on AI accounts that you connect: your own subscriptions and API keys. One page in Settings, AI, controls them, as on Squad for Mac. From top to bottom it has:

• Your squad’s AI: the model your squad thinks with, the model that reads images, and what it tries when it can’t answer.
• Accounts: your accounts on each provider.
• More: your squad’s defaults (How hard it thinks, Run time limit, and AI your squad may use). It is closed until you open it.

A teammate’s own model, heartbeat, and routines are on Setup on that teammate’s page. To connect your first account, see Connect your AI subscription. To open AI, click Settings at the bottom of the sidebar, then AI. You can also press ⌘K (Ctrl+K on Windows) and select AI, or select Squad defaults to open AI with More open. On a narrow window, the account menu also opens AI.

​

How Squad picks a model and an account

Every run follows the same order:

1. The run starts on the teammate’s own model, if it has one. Every other teammate uses the model in Thinks with.
2. On that model’s provider, Squad uses the accounts in the order on the provider’s card. An account that is expired, invalid, or at its limit is skipped.
3. When the model fails or hits a rate limit, the run moves to the next model under When it can’t answer.
4. Screenshots and photos go to the image model: the teammate’s own image model, or the model in Reads images with.

In History, open a run. Technical details shows the Model and Provider that ran. See See what your squad did.

​

Add another account for a provider

You can connect more than one account for each provider, for example two ChatGPT accounts. Only the owner, and an Administrator who works with every teammate, can add, replace, or disconnect accounts. Other people see the note “You can see which AI accounts the squad runs on. The squad owner or an Administrator who works with every teammate connects and changes them.” On Squad for Mac, the note also says that they can see how much the accounts have used. If the name is already in use, the dialog says that connecting would overwrite that account, and the connect button stays off. Pick another name, or use Replace on the existing account. Some providers have rules for extra accounts:

• Extra Qwen Cloud keys must come from the same plan.
• Extra Cloudflare Workers AI tokens must use the same account ID.
• MiniMax has no Token name field.
• Claude Code uses one active account at a time. See Set the order in which accounts are used.

​

Read a provider card

Under Accounts, each provider with at least one account has a card. From top to bottom, the card shows:

​

Read an account row

Each row shows the account name, a status word only when something needs attention, a detail line when there is more to say, usage meters when the provider reports them, and the … menu (Actions for this account): Set email or Edit email (Claude Code only), Replace…, and Disconnect…. A signed-out or expired account also has Reconnect on its row. The detail line shows the account’s email or plan, or a masked key. It can also show:

• for ChatGPT · Codex, only when a renewal failed: “It couldn’t renew itself. Reconnect it to sign in again.” See Keep ChatGPT accounts renewed.
• for Grok: expired, or, when only some teammates hold an expired copy, “expired on 3 of 9 teammates”.

Usage meters show for ChatGPT · Codex, Claude, Claude Code, OpenRouter, Kimi, Z.ai, and GitHub Copilot:

• Each limit window has a meter with the percentage used and the time until it resets, for example “62% used · resets in 3 hours”.
• A meter turns amber from 80% used and red from 95% used.
• Some providers show Balance: and the amount left.
• A thin bar shows while Squad checks the usage. “We couldn’t read this account’s usage right now.” means that the check failed. The account can still work.

Under the rows, the card can show one of these lines: Row buttons:

• Reconnect and Replace… both open the dialog Replace and the account name, and the new login overwrites exactly that account.
• Disconnect… removes the account after you confirm in the dialog “Disconnect [name]?”: “Teammates who use this account stop until you connect one again.” Squad refuses to disconnect the Claude Code account that your teammates use now: “This is the account your agents are using right now. Switch to another Claude account first, then disconnect this one.” Move another account to the top and save the order first. If a disconnect fails, the reason shows under the account’s row, for example “Couldn’t remove the [provider] connection from your squad.” and the cause.
• On Squad for Mac, the menu has no Replace… when the row shows Reconnect.
• Set email or Edit email shows on Claude Code rows only, because a Claude Code token does not show which account it belongs to. Type the email and click Save. To remove the email, clear the field and save.

​

Set the order in which accounts are used

On a card with two or more accounts, the order is the fallback order. Squad tries the top account first. When that account cannot serve a run, for example because it reached a usage limit, Squad tries the next one. For ChatGPT, this also covers an account whose plan cannot run the teammate’s model. When no ChatGPT account can run it, the run moves to the next model under When it can’t answer.

1. On the provider’s card, use the arrows Move this account up and Move this account down on the account rows.
2. In the bar that shows Priority order changed, click Save changes. Reset undoes the change.

The bar shows Saving the new account order… while it saves. Claude Code uses one account at a time, the active account. When the active account reaches its usage limit, the next account in the order becomes active, and the full account is skipped until its limit resets. The run that hit the usage limit continues once on the new account. If it had already done work, the teammate is told to continue from where it stopped and not to repeat what is done. A rate limit, when the provider slows down requests for a short time, does not switch Claude Code accounts. The run moves to the next model under When it can’t answer, if you set one. Otherwise it fails. The newest connection becomes the active account. When you save a new order, the first account becomes the active account. A Claude Code account that ran out shows At its limit until its limit resets. If no other Claude Code account can take over, Claude Code runs start again when the limit resets or when you connect another account.

​

Fix credentials that are out of sync

When an account does not match on every teammate, the top of AI shows “Some teammates hold an older copy of an account.”, and the account’s row shows Out of date for some teammates. When every teammate has the same credentials, the page shows nothing about it. Only the owner and Administrators see the warning.

1. Click Update them. It shows Updating…. If the update fails, the page says “Couldn’t update them. Try again.”
2. Wait until the warning goes away.

​

Keep ChatGPT accounts renewed

A ChatGPT · Codex account renews its sign-in by itself, and the row says nothing while it does. If a renewal fails, the row says “It couldn’t renew itself. Reconnect it to sign in again.” Select Replace… in its … menu and sign in again. When ChatGPT reports a banked limit reset for an account, the row shows “One banked limit reset is available on this account.” or the number of resets, and a Use one now button.

1. Click Use one now.
2. In Use one banked reset?, click Use one reset.

This refreshes the account’s usage limits now. You cannot get a spent reset back. The row then shows how many banked resets are left.

​

Choose the default model and fallback models

The card Your squad’s AI is at the top of AI, above Accounts. Everyone who opens AI sees it. Only the owner and an Administrator who works with every teammate can change it. The lists show only the models that your squad may use (AI your squad may use under More) and whose AI account is connected. Models are grouped by provider, with Claude Code first. The Anthropic group, for the Claude provider, is last and says “Deprecated: use Claude Code instead.” A GitHub Copilot model that your Copilot plan does not include is not offered in any list. On Squad for Mac, the lists also leave out models whose account is not connected, for every provider that has a card on AI. They can still offer a Copilot model that your plan does not include, and the models of a provider without a card on the Mac, such as Cloudflare Workers AI. To pick such a model, use the dashboard in a browser. Squad for Mac 1.0.98 does not offer Claude Code Fable 5.2 and Claude Fable 5.2 yet. To add them to your squad’s models, use the dashboard in a browser too. The last option in each list, Other models…, opens a list of the models that your squad cannot choose yet and whose account is connected, grouped by provider. On Squad for Mac, this list also shows providers whose account is not connected. It says “It joins the models your squad can use when you save.” The model that you pick fills that place, and it joins AI your squad may use when you click Save changes. Cancel closes the list. If the list has no fallbacks, the card says “Nothing set: when it can’t answer, the work waits until it can.” Add at least one fallback on a different provider. A model already picked whose account is not connected shows “[model] runs on [account], which isn’t connected, so its work fails. Pick a model from an account below.” A picked model that is not offered for another reason, for example a GitHub Copilot model that your cloud computer has not loaded yet, shows “[model] isn’t available right now, so it isn’t in the list. Choosing another model replaces it.” In When it can’t answer, such models show “Its account isn’t connected” or “It isn’t available right now”, and their Use first is off. On Squad for Mac, a fallback whose account is not connected shows “Its account isn’t connected”, and its Use first is off. If no model is saved, Thinks with shows Not set. If the image model cannot read images, the card says “[model] can’t read images, so image tasks fail on it. Pick a model that can.” If the card cannot load, it shows “We couldn’t read your squad’s AI just now.” Click Try again.

​

Give one teammate its own model

Most teammates use your squad’s AI. To let one teammate think with another model, or read images with another model, set it with Setup at the end of the teammate’s About page. See Settings reference and Add, change, and pause teammates.

​

Set your squad’s defaults under More

Your squad’s defaults are under More, at the bottom of Settings > AI. Click More to open or close it. The model that your squad thinks with and reads images with is set only on Your squad’s AI. The levels of How hard it thinks are Auto (per-model default), Off: disable reasoning, Minimal, Low, Medium, High, Extra high, Max (GPT-5.6 family), and Adaptive (model picks). Levels that your squad’s model does not support are not listed. A saved level that the model does not support shows · not supported, and Squad refuses to save until you pick another level. This level applies to the whole squad only. You cannot set it for one teammate or one routine. Routines have their own model field. See Schedule recurring work. AI your squad may use groups models by provider. Each group shows “N of M enabled” or None enabled. Click a group to open or close it, and click a model to allow or disallow it. Models from your plan that are not in the standard list show in a From your plan group. You can also allow a model with Other models… in a model list. See Choose the default model and fallback models. To save:

1. Change the settings.
2. In the bar that shows Unsaved changes, click Save changes. Reset undoes your edits.

Closing More keeps your unsaved changes and the bar. Nothing restarts, and runs in progress keep going. The next run uses the new settings. If your cloud computer has not applied the change yet, a notice says Saved. Applying changes in the background (usually a minute or two). until it has, at most for the time in Limits and defaults. If More cannot load, it shows “We couldn’t read your squad’s AI just now.” Click Try again. More can also show On your computer, not in your squad. See Add, change, and pause teammates.

​

What your teammates can do here

A teammate can change some of these settings for you when you ask:

• For anyone: list the AI accounts.
• For the owner and Administrators who work with every teammate: choose which account of a provider is used first, and change the email shown on an account.
• For the owner and Administrators: change one teammate’s model or image model. A model must be one that your squad may use.
• For the owner and Administrators who work with every teammate: change the squad’s default model and image model.

Squad gives teammates no tool to connect, replace, reconnect, or disconnect an AI account, to change the models under When it can’t answer, or to change AI your squad may use. Make these changes yourself on the pages that this page describes. See Your squad’s settings.

​

Check that it works

• Under Accounts, one account shows In use on each provider that reports usage.
• Your squad’s AI shows your model in Thinks with, with “Through” and the provider name, and your fallbacks under When it can’t answer.
• In History, open a run. Technical details shows the Model and Provider that ran.

​

Troubleshooting

Health says that a provider is rate-limiting your teammates. Cause: three or more runs in a row failed with rate-limit errors from that provider in the last six hours. Health shows it under Provider status. The provider’s card on AI shows Hit its limit, cooling down after one recent rate-limit failure. Fix: add another account for that provider, or add a model on a different provider under When it can’t answer. Squad does not start a run again after it failed with a rate-limit error. The only exception is the Claude Code account switch described in Set the order in which accounts are used. Check that the work finished. Send the request again, or wait for the next scheduled run, after the limit resets. Chat says “Your AI provider account is out of credits.” Cause: your provider refused the call because the account has no credits, balance, or plan left. Some providers answer this the way they answer a rate limit. Squad shows it as out of credits and does not try the message again. Fix: add credits or a plan at your provider, or connect another account and put it first in the card’s order. Then select Retry. See Chat failure messages. Chat shows “Your AI connection expired. Replies will fail until you reconnect.” Cause: the login for the model’s provider expired. Fix: click Reconnect, then click Reconnect on the account that shows Expired or Signed out. If its row shows neither, select Replace… in its … menu. For the owner and Administrators, a failed reply also shows Reconnect, which opens the same page. A reply says that a teammate’s model “isn’t available on your AI account”. Cause: your AI provider refused the model, for example because it no longer serves that model to your account. On ChatGPT, Squad first tries your other ChatGPT accounts and the models under When it can’t answer, so you see this only when none of them could answer. Fix: connect a ChatGPT account whose plan includes the model, or select Choose a model under the reply, or open Setup on the teammate’s page, and pick another model for the teammate. Then select Retry. In a group chat there is no Retry, so send the message again. See Chat failure messages. AI says that a provider “is signed out on your squad”. Cause: the provider’s login is no longer valid on your cloud computer. Runs that use it fail until you reconnect it. A banner above every other screen also says “[account] is signed out, so your squad can’t work.” For ChatGPT, the banner shows only when no ChatGPT account works. For Claude Code, the owner also gets one email. See An expired AI connection. Fix: click Reconnect on the account on its card. If its row shows neither Signed out nor Expired, select Replace… in its … menu. AI shows “This teammate is asleep”. Cause: the dashboard could not reach your cloud computer. Fix: click Wake & retry. A model is missing from a list. Cause: the model is not in AI your squad may use, its AI account is not connected, or your GitHub Copilot plan does not include it. Fix: pick it with Other models… in the list, or turn it on under More > AI your squad may use, then click Save changes. Or connect an account for its provider. On a teammate’s Setup, a model that a teammate still uses but that the lists do not offer shows why after its name, for example (not on your Copilot plan) or (not ready on your cloud computer yet). See Change a teammate’s model. A line says “Claude Code has reached its usage limit.” Cause: every Claude Code account that you connected reached its usage limit. It is not a sign-out. Fix: wait until the time that the line gives, or select Add another account. To keep work going during a limit, add a model under When it can’t answer. See Claude Code reached its usage limit. A reply says “No AI account is connected for [name]’s model.” Cause: no account that you connected can run the teammate’s model. Fix: connect an account for that model’s provider, or pick another model for the teammate. Then select Retry. A reply says that your Google AI key is on Google’s free plan. Cause: Google’s free plan does not include the model, or its allowance is used up for now. Squad does not try a model that the plan leaves out again, and moves to the next model at once. Fix: set up billing for the key in Google AI Studio, or pick another model. See Chat failure messages. An account shows At its limit, but you have usage left on the provider’s website. Cause: one of the account’s limit windows is full, or its balance is spent. Fix: wait for the reset time on the meter, or move another account to the top. A model you used is now called Kimi K2.8 Preview. Cause: Kimi now serves a newer model under the same name, and Squad renamed Kimi K2.7 Coding to match. Fix: nothing. Your settings keep it. A model you used is now called Qwen 3.8 Max (OpenCode). Cause: OpenCode replaced Qwen 3.7 Max with Qwen 3.8 Max. Squad moved your settings to the new model. Fix: nothing. You do not need to pick it again.

​

Common questions