Skip to content

Guides and concepts for the OpenAI API

Endpoints, parameters, and responses

Build apps to extend ChatGPT

Build commerce flows in ChatGPT

Assets for developers building with OpenAI

Notebook examples for building with OpenAI models

Learnings and experiences from developers

Search the API docs

Primary navigation

Get started

Overview
Quickstart
Models
Pricing
Libraries
Latest: GPT-5.2

Core concepts

Text generation
Code generation
Images and vision
Audio and speech
Structured output
Function calling
Responses API

Agents

Overview
Build agents
Deploy in your product
Optimize
- Agent evals
- Trace grading
Voice agents

Tools

Using tools
Connectors and MCP
Skills
Shell
Web search
Code interpreter
File search and retrieval
- File search
- Retrieval
More tools

Run and scale

Conversation state
Background mode
Streaming
WebSocket mode
Webhooks
File inputs
Context management
Prompting
- Overview
- Prompt engineering
Reasoning
- Reasoning models
- Reasoning best practices

Evaluation

Getting started
Working with evals
Prompt optimizer
External models
Best practices

Realtime API

Overview
Connect
Usage

Model optimization

Optimization cycle
Fine-tuning
Graders

Specialized models

Image generation
Video generation
Text to speech
Speech to text
Deep research
Embeddings
Moderation

Going live

Production best practices
Latency optimization
Cost optimization
Accuracy optimization
Safety

Legacy APIs

Assistants API

Resources

Terms and policies
Changelog
Your data
Permissions
Rate limits
Deprecations
MCP for deep research
Developer mode
ChatGPT Actions

Getting Started

Overview
Quickstart
Explore
Pricing
Concepts

Using Codex

App
IDE Extension
CLI
Web
Integrations

Configuration

Config File
Rules
AGENTS.md
MCP
Skills
Multi-agents

Administration

Authentication
Security
Enterprise
Windows

Automation

Non-interactive Mode
Codex SDK
App Server
MCP Server
GitHub Action

Learn

Videos
Blog
Cookbooks
Building AI Teams

Community

Ambassadors
Meetups

Releases

Changelog
Feature Maturity
Open Source

Home
Quickstart

Core Concepts

MCP Apps in ChatGPT
MCP Server
UX principles
UI guidelines

Plan

Research use cases
Define tools
Design components

Build

Set up your server
Build your ChatGPT UI
Authenticate users
Manage state
Monetize your app
Examples

Deploy

Deploy your app
Connect from ChatGPT
Test your integration
Submit your app

Guides

Optimize Metadata
Security & Privacy
Troubleshooting

Resources

Changelog
App submission guidelines
Reference

Home

Guides

Get started
Key concepts
Production readiness

Commerce specs

Agentic Checkout
Delegated Payment

Product feeds

Overview
Onboarding
Feed spec
Best practices

Home
Docs MCP

Categories

Code
Cookbooks
Guides
Videos

Topics

Agents
Audio & Voice
Computer use
Codex
Evals
gpt-oss
Fine-tuning
Image generation
Scaling
Tools
Video generation

Home

Topics

Agents
Evals
Multimodal
Text
Guardrails
Optimization
ChatGPT
Codex
gpt-oss

Contribute

Cookbook on GitHub

All posts

Recent

Building frontend UIs with Codex and Figma
Shell + Skills + Compaction: Tips for long-running agents that do real work
15 lessons learned building ChatGPT Apps
Testing Agent Skills Systematically with Evals
Supercharging Codex with JetBrains MCP at Skyscanner

Topics

General
API
Apps SDK
Audio
Codex

API API Reference Codex ChatGPT Learn

API Reference
API Reference
Responses API
Webhooks
- Events
- Unwrap
Platform APIs
- Audio
- Videos
- Images
- Embeddings
  - Create an embedding
- Evals
- Fine Tuning
  - Jobs
    Create a job
    Retrieve a job
    List jobs
    List Events
    Cancel a job
    Pause
    Resume
    Checkpoints
    List checkpoints
  - Checkpoints
    Permissions
    Create a permission
    Retrieve a permission
    Delete a permission
  - Alpha
    Graders
    Run
    Validate
- Batches
- Files
- Uploads
- Models
- Moderations
  - Create a moderation
Vector Stores
- Vector Stores
- Files
- File Batches
ChatKit
- Sessions
  - Create a session
  - Cancel a session
- Threads
Containers
- Containers
- Files
Skills
- Skills
Realtime
- Calls
- Client Secrets
  - Create a client secret
- Connect
- Client events
- Server events
Administration
- Overview
Chat Completions
- Chat Completions
Legacy
- Realtime Beta
  - Overview
- Assistants
- Completions
  - Create a completion

Input Tokens

Get input token counts

responses.input_tokens.count(InputTokenCountParams**kwargs) -> InputTokenCountResponse

POST/responses/input_tokens