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

# Models and Auto

> Use Auto, or choose a specific AI model for the job

Cleve works with models from OpenAI, Anthropic, Google, xAI, Meta, and leading open-model labs. You can switch models at any point in a conversation, and Cleve keeps the context.

## Auto (recommended)

**Auto** is the default model and the one we recommend for most work.

With Auto, Cleve chooses the model for you. It balances quality and speed against how much of your usage each reply takes, so your plan goes further. It handles everyday writing, research, planning, and editing well.

> [!TIP]
>
> Start with Auto. Switch to a specific model only when you want a particular model's style, or the strongest reasoning for an unusually hard task.

If you use Cleve very heavily in a single day, Auto may switch to a faster model for the rest of that day so your usage lasts longer.

## Choose a model

Click the model name in the message box to open the model picker, then pick a model. The new model answers from your next message onwards, with the whole conversation still in context.

Models your plan doesn't include show a lock. Selecting one offers an upgrade.

## Available models

| Provider | Models | Plans |
| - | - | - |
| **Cleve** | Auto | All plans |
| **Google** | Gemini 4 Argon, Gemini 3.8 Flash | All plans |
| **OpenAI** | GPT-OSS 120B | All plans |
| **OpenAI** | GPT-6.1 Sol, GPT-6 Luna | Pro and Max |
| **Anthropic** | Claude Sonnet 5.5, Claude Haiku 5.5 | Pro and Max |
| **xAI** | Grok 4.7 | Pro and Max |
| **DeepSeek** | DeepSeek V4 Pro, DeepSeek V4.1 Flash | Pro and Max |
| **Z.ai** | GLM 5.3 | Pro and Max |
| **Anthropic** | Claude Opus 5.5 | Max |
| **Meta** | Muse Wonder | Max |
| **Moonshot** | Kimi K3 | Max |

Team plans include models by seat type. See the [pricing page](https://www.cleve.ai/pricing) for the current details of each plan.

### Open models

Kimi K3, DeepSeek V4 Pro, DeepSeek V4.1 Flash, and GLM 5.3 appear under **Open models (US-provider hosted)** in the picker. They run on infrastructure from US-based providers.

### Legacy models

Previous-generation models stay available under a collapsed **Legacy models** section, so older conversations keep working. For new work, use Auto or a current model.

## Which model should I use?

| If you want… | Try |
| - | - |
| A great default for almost everything | **Auto** |
| Fast replies for quick tasks | Gemini 3.8 Flash, GPT-6 Luna, Claude Haiku 5.5, DeepSeek V4.1 Flash |
| Strong, polished writing | Claude Sonnet 5.5, GPT-6.1 Sol |
| The deepest reasoning on hard problems | Claude Opus 5.5, Muse Wonder, Kimi K3 |
| Code and technical work | DeepSeek V4 Pro, GPT-OSS 120B |

More capable models use your usage faster. Auto gets you the most out of your plan. See [How usage works](https://docs.cleve.ai/pricing/usage.md).

## Privacy

Every model in Cleve runs with **zero data retention**: AI providers don't keep your prompts or outputs after responding. Your docs and chats are never used to train AI models.


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