Skip to main content

API Platform

The most powerful platform for building AI products

Build and scale AI experiences powered by industry-leading models and tools.

Start building(opens in a new window)View API pricing

Flagship Models

Our reasoning models

These models spend more time thinking before producing a response, making them ideal for complex, multi-step problems.

OpenAI o1

Our frontier reasoning model that supports tools, Structured Outputs, and vision

  • Text and vision

  • 200k context length

  • Input: $15 | Output: $60 per 1M tokens

Learn more
OpenAI o3-mini

Our cost-efficient reasoning model that’s optimized for coding, math, and science, and supports tools and Structured Outputs

  • Text

  • 200k context length

  • Input: $1.10 | Output: $4.40 per 1M tokens

Learn more

Our GPT models

These general purpose models are designed to perform a wide variety of everyday tasks.

GPT-4o

Our high-intelligence model for complex tasks

  • Text and vision

  • 128k context length

  • Input: $2.50 | Output: $10
per 1M tokens

Learn more
GPT-4o mini

Our affordable small model for fast, lightweight tasks

  • Text and vision

  • 128k context length

  • Input: $0.15 | Output: $0.60 per 1M tokens

Learn more

Access the power of our models with APIs

Chat Completions API

Get access to our most powerful models with a few lines of code.

Learn more
A screenshot of code showing a request to an AI model (`gpt-4.0`) to create a haiku. The displayed haiku reads: "Silent circuits hum, Echoes of thought, code woven— Dreams in silicon." The code uses a `completion` function to generate the text.

Build AI-native experiences with our tools and capabilities

Knowledge retrieval

Give the model access to your data for intelligent retrieval in your AI applications. 

Learn more
The image shows a file management interface for "Quarterly earnings." It includes details like file size, last active date, expiration policy, and creation date. Three PDF files are attached, with options to add files, create an assistant, or copy IDs.

Code interpreter

Get models to run code iteratively to solve challenging code and math problems, and generate charts.

Learn more
The image shows a Python code snippet in a math tutorial explaining how to generate the first 10 Fibonacci numbers and calculate their sum. The code appends the Fibonacci numbers to a list and computes the sum, the output showing the sequence and total.

Function calling

Instruct the model to intelligently interact with your codebase and APIs using custom functions.

Learn more
The image shows a Python code snippet where an assistant is created using OpenAI's API. The assistant is programmed as a weather bot with functions to retrieve the current temperature and the probability of rain for a specified location using GPT-4o.

Vision

Get the model to understand and answer questions about images using vision capabilities.

Learn more
The image shows an AI chat where a user asks, "What is this plant?" with a small plant image. The assistant identifies it as Sansevieria trifasciata, providing details about its care and uses. Settings for temperature and privacy are on the right.

JSON Mode

Guarantee JSON outputs from the model when you enable JSON mode.

Learn more
The image shows a cURL command using OpenAI's API to send a request for the population of the three largest cities in the world. The assistant responds in JSON format, listing Tokyo, Delhi, and Shanghai with their respective populations.

Streaming

Display model outputs in real-time as they are generated.

Learn more
A user typing "How do I find the Milky Way at night?" and the GPT displays model outputs in real-time as they are generated.

Customize models for your needs

Fine-tuning

Customize a model’s existing knowledge and behavior for a specific task using text and images via supervised fine-tuning. Plus, use our Model Distillation tools to fine-tune smaller models on the outputs of more capable models.

Learn more
The image shows a model training interface for "ft:gpt-6." It includes details like status, base model, creation date, checkpoints, and steps. A graph at the bottom shows the training and validation loss over time, showing the model's performance progress.

See what you can build in Playground

Explore our models and APIs in Playground without writing a single line of code.

Start exploring
The image shows a chat where a user asks, "How do I get started with OpenAI's API?" The assistant provides steps on signing up, getting API keys, and accessing documentation. Temperature and token settings are visible on the right.

Trusted by the world’s most innovative organizations

Oscar Cover Image

Oscar brings AI to health insurance, reducing costs and improving patient care

Story

Black Harvey logo on a white background.

Harvey partners with OpenAI to build a custom-trained model for legal professionals.

Story

Screenshot 2024 03 12 At 1128 27am

Superhuman introduces a new era of email with OpenAI.

Superhuman

Healthify Open Ai Cover Image 1

Healthify collaborates with OpenAI to improve millions of lives with sustainable weight loss.

Healthify

Enterprise-grade features for operating at scale

Security and data privacy

Administrative controls

  • Scope user roles and API keys to individual projects

  • Role-based access controls

  • Limit access to specific models within each project

  • Set billing and usage restrictions to avoid overages

  • View granular usage activity by project

Access to OpenAI expertise

  • Dedicated account team and prioritized support

  • Guidance on AI deployment and best practices from solutions architects

  • Opportunity to collaborate with researchers on building custom AI models

Three million developers build with OpenAI

Sign up for a developer account(opens in a new window)Talk to our team