# o1

> OpenAI's reinforcement-learning-trained reasoning model for complex problems, with image input and tools.

## Key facts

- Model author: OpenAI
- Input: $15.00 per 1M tokens
- Output: $60.00 per 1M tokens
- Cached input: $7.50 per 1M tokens
- Context window: 200,000 tokens
- Maximum output: 100,000 tokens
- Input modalities: image, text
- Output modalities: text
- Capabilities: reasoning, structured_output, tools, vision
- Released: 2024-12-17
- Knowledge cutoff: Oct 01, 2023
- Model IDs: `openai/o1`
- Model page: https://www.realrelay.ai/model/openai/o1

## Overview

o1 is a reasoning model from OpenAI. OpenAI documents the o1 series as trained with reinforcement learning to perform complex reasoning, and the model is designed to spend more time thinking before it responds. It accepts text and image input, returns text, and supports tool calling and structured output.

It has a 200,000-token context window and can generate up to 100,000 output tokens. It accepts text and image input and returns text. Its knowledge cutoff is October 1, 2023, and it was released on December 17, 2024.

OpenAI lists o1 as deprecated in its model documentation.

## Capabilities

o1 supports reasoning, tool calling, structured output and image understanding. It reads text and image input together and returns text.

Its defining trait is deliberate reasoning: it works through a problem before answering, which suits tasks where accuracy on a multi-step question matters more than a quick reply.

## Pricing

o1 is priced at $15.00 per 1M input tokens and at $60.00 per 1M output tokens. Cached input reads are priced at $7.50 per 1M tokens.

o1 is available through RealRelay's OpenAI-compatible API.

## Use cases

Use o1 for multi-step problems that reward careful thinking, such as analyzing a technical document, tracing a bug through unfamiliar code or working through a calculation. The 200,000-token context window holds long inputs.

Image input lets it reason over diagrams, charts and screenshots, and structured output returns the answer in a form other code can use.

## 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/o1", "messages": [{"role": "user", "content": "Hello"}]}'
```

## FAQ

### What is o1?

o1 is an OpenAI reasoning model trained with reinforcement learning for complex reasoning, with image input, tool calling and a 200,000-token context window. OpenAI lists it as deprecated.

### How much does o1 cost?

o1 is priced at $15.00 per 1M input tokens and at $60.00 per 1M output tokens. Cached input reads are priced at $7.50 per 1M tokens.

### What is the context window and maximum output of o1?

o1 has a 200,000-token context window and can generate up to 100,000 output tokens.

### What inputs and outputs does o1 support?

o1 accepts text and image input and returns text.

### What capabilities does o1 support?

o1 supports reasoning, tool calling, structured output and image understanding.

### When was o1 released?

o1 was released on December 17, 2024.

### How do I call o1 via the API?

Call o1 through RealRelay's OpenAI-compatible API using the model ID openai/o1.