# GPT-4

> An OpenAI text model that takes text input and returns text, served through RealRelay's OpenAI-compatible endpoint.

## Key facts

- Model author: OpenAI
- Input: $30.00 per 1M tokens
- Output: $60.00 per 1M tokens
- Context window: 8,192 tokens
- Maximum output: 8,192 tokens
- Input modalities: text
- Output modalities: text
- Capabilities: fine_tuning
- Released: 2023-05-28
- Knowledge cutoff: Dec 01, 2023
- Model IDs: `openai/gpt-4`
- Model page: https://www.realrelay.ai/model/openai/gpt-4

## Overview

GPT-4 is an OpenAI text model offered on RealRelay under the model ID openai/gpt-4. Requests carry text input and receive text output, and calls use RealRelay's OpenAI-compatible endpoint.

The context window is 8,192 tokens and the maximum output is also 8,192 tokens. The model was released on May 28, 2023, and its knowledge cutoff is December 1, 2023. Fine-tuning is listed among its capabilities.

## Capabilities

Input and output are both text, so a request passes text and receives text. Fine-tuning is listed as a supported capability of the model.

Within one request the context window is 8,192 tokens and the maximum output length is 8,192 tokens.

## Pricing

Pricing is metered by token. Input is $30.00 per 1M tokens and output is $60.00 per 1M tokens.

Each call made through RealRelay's OpenAI-compatible endpoint with the model ID openai/gpt-4 is billed at those token rates.

## Use Cases

Because input and output are both text and the context window is 8,192 tokens, the model suits text requests whose prompt and completion both fit inside that window.

When prompting alone is not enough, the fine-tuning capability listed for the model gives a way to adapt its behaviour to your own data.

To call it, set the model to openai/gpt-4 in an OpenAI client and point the client at RealRelay.

## Call it

OpenAI-compatible chat completions:

```bash
curl https://www.realrelay.ai/v1/chat/completions \
  -H "Authorization: Bearer $REALRELAY_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{"model": "openai/gpt-4", "messages": [{"role": "user", "content": "Hello"}]}'
```

## FAQ

### What is GPT-4?

GPT-4 is an OpenAI text model that accepts text input and returns text output. On RealRelay it is called with the model ID openai/gpt-4.

### How is GPT-4 priced?

Input is $30.00 per 1M tokens and output is $60.00 per 1M tokens.

### What is the context window of GPT-4?

The context window is 8,192 tokens, and the maximum output is also 8,192 tokens.

### Which input and output types does GPT-4 support?

GPT-4 accepts text input and returns text output.

### Does GPT-4 support fine-tuning?

Fine-tuning is listed among the model's capabilities.

### When was GPT-4 released?

GPT-4 was released on May 28, 2023, with a knowledge cutoff of December 1, 2023.

### How do I call GPT-4 through the OpenAI-compatible API?

Send requests to RealRelay's OpenAI-compatible endpoint with the model set to openai/gpt-4.