> ## Documentation Index
> Fetch the complete documentation index at: https://docs.chat402.xyz/llms.txt
> Use this file to discover all available pages before exploring further.

# Using the Chat Interface

> How to chat with AI models on Chat402

## Chatting with AI Models

Select from premium AI models and start chatting instantly.

### Selecting a Model

Available models in the tabs:

* **GPT-5.2** - OpenAI's best model (\$0.020/1K tokens)
* **GPT-3.5 Turbo** - Fast, general-purpose (\$0.002/1K tokens)
* **GPT-4 Turbo** - Complex reasoning (\$0.040/1K tokens)
* **Claude 4.5 Sonnet** - Analysis, long-form (\$0.018/1K tokens)
* **Gemini 2.5 Flash** - Cost-effective (\$0.003/1K tokens)
* **Grok** - xAI's latest (\$0.001/1K tokens)
* **DeepSeek Chat** - Code generation (\$0.001/1K tokens)
* **Synth** - Query all models at once (combined pricing)

### Sending Messages

1. Type your message in the input box
2. Press **Send** or hit **Enter**
3. See the AI's response with cost breakdown
4. Continue the conversation

### Cost Tracking

Each response shows:

* Token count (input/output)
* Total cost in USD
* Model used

Your total balance is displayed in the header.

## Special Features

### Synthesizer Mode

The **Synth** tab queries multiple AI models simultaneously and synthesizes their responses into one comprehensive answer.

**How it works:**

1. Your question is sent to multiple models (GPT-4, Claude, Gemini, etc.)
2. All responses are combined and synthesized
3. You get the best combined answer from all models

**Best for:**

* Complex questions that benefit from multiple perspectives
* Research and analysis tasks
* Getting more comprehensive answers

### LaTeX / Math Rendering

Chat402 supports LaTeX math notation. AI responses containing mathematical formulas are automatically rendered.

**Supported formats:**

* Inline math: `$E = mc^2$` renders as inline formula
* Display math: `$$\int_0^\infty e^{-x^2} dx = \frac{\sqrt{\pi}}{2}$$` renders as block formula
* Alternative: `\(inline\)` and `\[display\]` also work

**Use cases:**

* Ask for mathematical explanations
* Get step-by-step equation solutions
* Physics and engineering calculations
* Statistics and probability formulas

### Markdown Support

Responses are rendered with full Markdown support:

* **Bold** and *italic* text
* Code blocks with syntax highlighting
* Tables
* Lists and numbered items
* Links and images

### Image Uploads

For vision-capable models (GPT-4 Vision):

1. Click the image upload button
2. Select an image file
3. Ask questions about the image
4. Get detailed visual analysis
