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

# Model Selection Guide

> Which models Sixth supports, how to switch, and when to bring your own key.

Sixth works with the leading frontier models. On a Sixth plan they all run through your Sixth account, so you can switch freely without setting up API keys.

## Switching models

Click the model name in the bottom row of the message box (next to the **Act** mode button). A picker opens above the message box:

* **Search** by name, or browse models grouped by provider.
* Models marked **New** were added recently; **Popular** marks what most people use. The check mark shows your current model.
* Use `↑` `↓` to move, `Enter` to pick, `Esc` to close.
* The **Image** tab lists the image models Sixth can use for [image generation](/features/image-generation).

Your choice applies to new requests straight away, including in the middle of a task.

## Models available

| Provider | Models |
| - | - |
| Anthropic | Claude Opus 5.5, Fable 5.1, Fable 5, Claude Opus 4.8 / 4.7 / 4.6, Claude Sonnet 4.6, Claude Haiku 4.5 |
| OpenAI | GPT-6 Astra, GPT-6 Sol, GPT-6 Luna, GPT-5.6 Sol / Terra / Luna, GPT-5.5 |
| Google | Gemini 3.8 Flash, Gemini 3.1 Pro, Gemini 3.7 Flash, Gemini 3.5 Flash |
| xAI | Grok 4.7, Grok 4.6, Grok 4.5, Grok Build |
| Moonshot AI | Kimi K3 |
| DeepSeek | DeepSeek V4.1 Flash, DeepSeek V4 Pro |
| Meta | Muse Spark 1.3, Muse Spark 1.2 |
| Alibaba | Qwen 3.7 Plus, Qwen 3.6 27B, Qwen 3.6 Flash |
| Mistral AI | Mistral Medium, Mistral Small |

More options (OpenRouter, Hugging Face, Cerebras, SambaNova and local models via Ollama or LM Studio) appear further down the list. The picker always shows the current set, so it's the source of truth if this table falls behind.

## Extended thinking

Models that support extended thinking (shown with a lightbulb) let you turn reasoning on when you pick them. Choosing one opens a small panel with a **thinking** switch; turn it on and click **Use** to select the model with thinking enabled. Some models, such as Fable 5.1, always think.

While a model is thinking you'll see a live **Thinking** row in the chat; afterwards it collapses to **Thought for Ns**, and you can click it to read the reasoning.

## Choosing a model

* **Hard, multi-step coding work:** a top-tier model such as Claude Opus 5.5, Fable 5.1, GPT-6 Astra or Gemini 3.1 Pro.
* **Fast everyday edits:** a quicker model such as Gemini 3.8 Flash, DeepSeek V4.1 Flash, Claude Haiku 4.5 or Kimi K3.
* **Planning:** pair [Plan mode](/features/plan-and-act) with a strong reasoning model and turn thinking on.

You can switch at any point, so it's easy to plan with one model and implement with another.

## Context window

Each model has a context window: how much of the conversation, files and tool output it can hold at once. The ring next to the task title shows how much of it the current task is using. When a long task gets close to the limit, Sixth compacts the conversation automatically so you can keep going. See [Understanding context management](/getting-started/understanding-context-management).

## Bring your own key

If you'd rather use your own provider account, pick a model whose provider isn't set up yet. It shows an **Add key** badge, and selecting it opens an inline form for your API key (with a link to get one, plus a base URL field where relevant). Click **Save & use** and the model runs on your key from then on.

You can manage provider keys later in **Settings**. Setup guides for each provider are under [Provider Configuration](/provider-config/anthropic), and local models are covered in [Running models locally](/running-models-locally/read-me-first).


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