> ## Documentation Index
> Fetch the complete documentation index at: https://docs.aihubmix.com/llms.txt
> Use this file to discover all available pages before exploring further.

# Model FAQ

> Common questions about Harnesses, model sources, and reasoning effort: what is used by default, why models are grayed out, reaching limits, key failures, and whether switching models in a session loses context.

This page collects common questions about models. For details, follow the link in each answer.

## Choosing and defaults

### Which Harness and model does a new session use by default?

It reuses the choice from the last session you created. The first time, it is Claude Code + **Default model**. See [Default selection](https://docs.aihubmix.com/en/cloud/models/harness-and-models.md#default-selection).

### How do Plus models, your subscriptions, and BYOK differ?

Plus models come with the Plus plan, with no extra charge. Your subscriptions use the plan limits of your own subscription accounts. BYOK uses your own API keys, billed by the provider. See [Three model sources](https://docs.aihubmix.com/en/cloud/models/model-sources.md#three-sources).

### Where are Plus models? How do I change which models appear?

They are the first group in the BYOK tab of the model picker, and the first row of the **BYOK** section in settings. Click the pencil icon on this row to select which models appear. See [Where Plus models are](https://docs.aihubmix.com/en/cloud/models/model-sources.md#where-plus-models) and [Choose Plus models](https://docs.aihubmix.com/en/cloud/models/model-sources.md#choose-plus-models).

### How should I choose the reasoning effort?

When unsure, use the level marked **Recommended**. Use **High** for complex changes and **Extra high** for cross-file refactors. **Max** and Ultra use noticeably more, so use them only for the hardest tasks. See [Reasoning effort](https://docs.aihubmix.com/en/cloud/models/harness-and-models.md#reasoning-effort).

### Why is the effort menu sometimes a list and sometimes a slider?

The style of the effort menu follows the model and Harness: a level list or a slider, both labeled **Effort**. When the model does not support adjusting effort, the effort button is hidden. See [Two styles](https://docs.aihubmix.com/en/cloud/models/harness-and-models.md#effort-styles).

### How do Ultra and Ultracode differ?

Ultra is the highest level for some models in Codex CLI, using about 2–3 times as much as **Extra high**. Ultracode is a separate Claude Code toggle that splits a task into multiple subtasks run in parallel. See [Ultra level](https://docs.aihubmix.com/en/cloud/models/harness-and-models.md#ultra-effort) and [Ultracode toggle](https://docs.aihubmix.com/en/cloud/models/harness-and-models.md#ultracode).

### Why is the Ultracode toggle grayed out?

The current conditions are not met, and the reason is shown next to the toggle, for example the Harness is not Claude Code or the cloud server is not ready yet. See [Ultracode toggle](https://docs.aihubmix.com/en/cloud/models/harness-and-models.md#ultracode).

### Why did the reasoning effort change after I switched models?

When the new model does not have the level you chose, it returns to the new model's recommended level. See [How effort changes after switching models](https://docs.aihubmix.com/en/cloud/models/harness-and-models.md#effort-after-switch).

## Grayed out and unavailable

### Why are Plus models grayed out?

You do not have Plus yet. Click **Get Plus** below the group. See [Without Plus](https://docs.aihubmix.com/en/cloud/models/harness-and-models.md#without-plus).

### Why are the models of an account or key grayed out?

It is not compatible with the current Harness, and the reason is shown below the group. In a new session, clicking one switches to a compatible Harness automatically. In a session, they cannot be selected. See [Model picker](https://docs.aihubmix.com/en/cloud/models/harness-and-models.md#model-picker).

### My subscription is connected. Why are there no models in the Subscriptions tab?

Each account shows the reason, for example the current Harness does not support the subscription. Switch to another Harness or try again later. See [Model picker](https://docs.aihubmix.com/en/cloud/models/harness-and-models.md#model-picker).

### Why can't I use my Anthropic key with Codex CLI?

Codex CLI needs an **OpenAI compatible** API, and Claude Code needs an **Anthropic compatible** API. See [API format and Harness](https://docs.aihubmix.com/en/cloud/models/model-sources.md#api-format-harness).

### Why does adding a subscription say my cloud server isn't running?

Subscription sign-in happens on your cloud server, so the cloud server needs to be running. See [Add a subscription account](https://docs.aihubmix.com/en/cloud/models/model-sources.md#add-subscription).

### What do I do when a subscription account shows "Authorization expired" or "Signed in from a terminal"?

Click **Reconnect** on the account row. See [Reconnection needed](https://docs.aihubmix.com/en/cloud/models/model-sources.md#reconnect).

## Limits and failures

### What do I do when my subscription's plan limits are reached?

The session pauses, and changes and the environment are kept. You can continue with another account from the same provider, with another model, or after adding a new account. AIHubMix Cloud does not switch accounts automatically. See [Plan limits reached](https://docs.aihubmix.com/en/cloud/models/model-sources.md#subscription-limit-reached).

### Where can I see Plus limits?

The Plus models row shows a status when a limit is reached or Plus is temporarily unavailable. It does not show amounts or usage numbers. See [Plus limit status](https://docs.aihubmix.com/en/cloud/models/model-sources.md#plus-limit-status).

### What happens when a call with my API key fails?

By default, the turn stops with an error. With **Fall back to Plus models** on, it continues on Plus models. The toggle is unavailable until you add a key. See [Fall back to Plus models when a key fails](https://docs.aihubmix.com/en/cloud/models/model-sources.md#key-fallback).

### How do I replace an API key?

Click the pencil icon on the key row, enter the new key, and save. Leave it empty to keep the original key. See [Edit and delete keys](https://docs.aihubmix.com/en/cloud/models/model-sources.md#edit-delete-key).

## Switching models and context

### Can I switch models after a session starts? Can I switch Harnesses?

You can switch the model, source, and reasoning effort (after the current turn ends). You cannot switch the Harness. See [Harness is fixed after the session starts](https://docs.aihubmix.com/en/cloud/models/harness-and-models.md#harness-locked).

### Does switching models in a session lose context?

Usually not: all earlier records are kept. Only when the context does not fit in the new model are you asked to confirm compaction, and earlier details may no longer be kept. See [Switch models in a session](https://docs.aihubmix.com/en/cloud/models/harness-and-models.md#switch-model) and [Compact and switch](https://docs.aihubmix.com/en/cloud/models/harness-and-models.md#compact-and-switch).

### How do I know how much context the current session uses?

Hover the small ring to the left of the model button. See [Context usage ring](https://docs.aihubmix.com/en/cloud/models/harness-and-models.md#context-usage).

### If I change the models of Plus models or a key, are sessions in progress affected?

No. See [Notes and limitations](https://docs.aihubmix.com/en/cloud/models/model-sources.md#limitations).

## Other

### How is an AIHubMix API key related to Plus?

They are unrelated. AIHubMix is a regular gateway provider in BYOK, billed per key, and does not count toward Plus. See [Three model sources](https://docs.aihubmix.com/en/cloud/models/model-sources.md#three-sources).

### Can I connect multiple subscription accounts from the same provider?

Yes. Each account has its own name, and you can set one as the default. See [Multiple accounts and account names](https://docs.aihubmix.com/en/cloud/models/model-sources.md#multiple-accounts).

### What is the AI assistant? Do I need to change it?

It is a background model for each model source that handles background tasks such as resolving merge conflicts. The default is **Auto**, and usually you do not need to change it. See [AI assistant](https://docs.aihubmix.com/en/cloud/models/model-sources.md#ai-assistant).

### Which model does environment repair and chat titling use?

Check **Settings › Account › System task models**. See [System task models](https://docs.aihubmix.com/en/cloud/models/model-sources.md#system-task-model).

### Is there a shortcut to open the model picker?

Yes: `⌃⇧M` (`Ctrl+Shift+M` on Windows). See [Keyboard shortcuts](https://docs.aihubmix.com/en/cloud/models/harness-and-models.md#keyboard).

## Related

* [Harness and models](https://docs.aihubmix.com/en/cloud/models/harness-and-models.md)
* [Model sources: Plus models, your subscriptions, and BYOK](https://docs.aihubmix.com/en/cloud/models/model-sources.md)


This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.