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

# Model Picker

> How to browse, search, select, pin, and configure AI models using ZeroTwo's model picker.

The model picker lets you switch AI models at any time — before starting a conversation or mid-way through one. With 95+ models available from 19 providers, the picker gives you a fast, searchable interface to find exactly the model you need.

## Opening the Picker

The current model name is displayed as a clickable button on the **left side of the prompt bar** at the bottom of the chat interface.

<Steps>
  <Step title="Click the model name in the prompt bar">
    Find the model name button in the bottom-left of the prompt bar. It shows the currently active model (for example, "Claude Sonnet 4.6" or "GPT-5"). Click it to open the model picker panel.
  </Step>

  <Step title="The picker opens as a panel or modal">
    The picker slides open with a full, scrollable list of available models. Each model appears as a card showing the model name, provider, and tier (Standard or Premium).
  </Step>

  <Step title="Browse, filter, or search">
    Use the search bar at the top of the picker to filter by model name (type "claude", "gpt-5", "gemini", etc.). Use the provider sidebar on the left to filter by provider. Use the type tabs to switch between Text and Image model views.
  </Step>

  <Step title="Click a model to select it">
    Click any model card to select it. The picker closes, the model name in the prompt bar updates to reflect your selection, and all subsequent messages in this conversation will use the chosen model.
  </Step>
</Steps>

<Info>
  Switching models mid-conversation applies to new messages only. Prior messages in the conversation remain unchanged, and the full conversation history is sent to the new model as context.
</Info>

## Picker Interface

### Search Bar

Type any part of a model name to filter the list in real time. Examples: "claude", "gemini pro", "gpt", "flux", "deepseek". The search covers model names and provider names.

### Provider Filter

The left sidebar of the picker lists all 19 providers. Click any provider name to show only that provider's models:

* All (default)
* OpenAI
* Anthropic
* Google
* Mistral
* DeepSeek
* Cohere
* xAI
* Perplexity
* Qwen
* Groq
* OpenRouter
* Kimi K2
* Venice
* TheSys
* ZAI
* Inception
* ByteDance
* And more

### Type Filter

Tabs at the top of the picker switch between model types:

* **Text** — all chat and reasoning models
* **Image** — image generation models (also available when using the Image tool or navigating to `/studio/images`)

### Standard vs. Premium Badge

Each model card displays a badge:

| Badge        | Meaning                                                                                                     |
| ------------ | ----------------------------------------------------------------------------------------------------------- |
| **Standard** | Does not count against your premium message quota. Available in unlimited quantities on all paid plans.     |
| **Premium**  | Counts against your monthly premium message quota (e.g., 555/month on Pro). Not available on the Free plan. |

When your premium quota is exhausted, ZeroTwo automatically uses a fallback standard model. You can check your remaining premium quota in Settings → Account.

### Capabilities Indicators

Each model card shows key capability indicators, visible by hovering over the info icon or expanding the card:

| Indicator                | Meaning                                                                                                      |
| ------------------------ | ------------------------------------------------------------------------------------------------------------ |
| **Context window**       | Maximum tokens the model can process at once (e.g., 200k for Claude Sonnet 4.6, 1M for Gemini 2.5 Pro)       |
| **Vision support**       | Whether the model can analyze images attached to messages                                                    |
| **Tool use**             | Whether the model supports function calling / tool use (required for agent mode, web search, and connectors) |
| **Reasoning / Thinking** | Whether the model supports the thinking level slider for extended reasoning                                  |

### Common Context Windows

| Model                       | Context Window   | Approximate Length |
| --------------------------- | ---------------- | ------------------ |
| Gemini 2.5 Pro              | 1,000,000 tokens | \~750,000 words    |
| Claude Sonnet 4.6, Opus 4.6 | 200,000 tokens   | \~150,000 words    |
| GPT-4o                      | 128,000 tokens   | \~96,000 words     |
| GPT-5                       | 128,000 tokens   | \~96,000 words     |
| Mistral Large               | 128,000 tokens   | \~96,000 words     |
| DeepSeek Chat               | 64,000 tokens    | \~48,000 words     |

## Selecting a Model

<Steps>
  <Step title="Click the model name in the prompt bar">
    Opens the picker panel.
  </Step>

  <Step title="Optional: filter by provider or type in the search bar">
    Narrow the list to find the model you want faster.
  </Step>

  <Step title="Click the model card to select it">
    The picker closes and the model name in the prompt bar updates.
  </Step>

  <Step title="Your next message uses the selected model">
    Send your message as usual — ZeroTwo routes it to the newly selected model.
  </Step>
</Steps>

<Info>
  Changing the model in one chat does not affect any other chat. Each conversation maintains its own model selection independently.
</Info>

## Pinning Favorite Models

If you use certain models frequently, pin them to the top of the picker for instant access without scrolling.

<Steps>
  <Step title="Hover over a model card in the picker">
    Move your cursor over the model card to reveal the action icons on the right side of the card.
  </Step>

  <Step title="Click the star icon">
    Click the star (or pin) icon. The model is now pinned and will appear in a **Favorites** section at the very top of the picker, above all other categories.
  </Step>

  <Step title="Access pinned models instantly">
    The next time you open the picker, your pinned models appear at the top of the list — no scrolling or filtering needed.
  </Step>

  <Step title="Unpin when no longer needed">
    Click the star icon again on any pinned model to remove it from Favorites. It returns to its normal position in the full list.
  </Step>
</Steps>

<Tip>
  Pin 2–3 models you use daily — for example, a premium reasoning model for deep work and a fast standard model for quick tasks — to make switching between them nearly instant.
</Tip>

## Reasoning Models and the Thinking Level Slider

Several models in ZeroTwo support **extended reasoning** — they work through a problem step-by-step before producing a final response. When you select a reasoning-capable model, a **thinking level slider** appears in the prompt bar.

**Models with reasoning / thinking support:**

* OpenAI o3, o4-mini
* DeepSeek Reasoner
* Claude Sonnet 4.6 and Opus 4.6 (with extended thinking enabled)

### Thinking Levels

| Level      | Behavior                                    | Best For                                            |
| ---------- | ------------------------------------------- | --------------------------------------------------- |
| **Low**    | Minimal internal reasoning, faster response | Simple questions, quick tasks, iteration            |
| **Medium** | Moderate reasoning depth, balanced speed    | Most everyday tasks                                 |
| **High**   | Deep, thorough step-by-step reasoning       | Complex math, algorithmic coding, detailed analysis |

<Note>
  Higher thinking levels consume more tokens and result in longer response times. On plans with premium quotas, a "High" reasoning response uses more of your monthly quota than a standard response.
</Note>

## Switching Models Mid-Conversation

You can switch models at any point in an active conversation without losing any context:

1. Click the model name in the prompt bar
2. Select a different model from the picker
3. Send your next message — it is handled by the new model

The new model receives the full conversation history up to that point. This means you can:

* Start with a fast standard model for early brainstorming
* Switch to a reasoning model (o3, DeepSeek Reasoner) when the problem gets complex
* Switch to Perplexity Sonar when you need live web-grounded answers
* Switch to a large-context model (Gemini 2.5 Pro) when pasting in a large document

<Warning>
  Not all models support all tools. If you switch to a model that does not support tool use, active tools like Web Search and Agent Mode will be disabled for that session. Check the capabilities indicators before switching if you rely on specific tools.
</Warning>

## Image Model Selection

When you enable image generation — either via the Image tool pill in the prompt bar or by navigating to `/studio/images` — the picker updates to show available image generation models:

* GPT-Image-1, GPT-Image-1.5, GPT-Image-mini
* Flux Pro v1.1, Flux Pro v2
* Imagen 4.0
* Grok Imagine, Grok Imagine Pro
* Qwen Image, Qwen Edit
* LustIFY SDXL and other models

Each image model has different stylistic characteristics, resolution options, and latency profiles. See [Models Overview](/overview/models-overview) for a full comparison of image model strengths.

## Related Pages

* [Models Overview](/overview/models-overview) — full list of all text, image, video, and audio models organized by provider
* [Plans and Pricing](/overview/plans-and-pricing) — which models count as premium and how quotas work
* [Answer Quality and Limitations](/core-chat/answer-quality-and-limitations) — guidance on choosing models for specific task types
* [Chat Interface Overview](/core-chat/overview) — full tour of the prompt bar and tool pills