This page is the single reference document for everything ChatGPT can do. Capabilities, plans, platform availability, security posture, and how to start using it — all in one place. Bookmark it. Come back when you need the definitive answer on any ChatGPT question.
ChatGPT reached 200 million weekly active users faster than any software product in history. Below is the complete picture of why.
ChatGPT is an AI chatbot developed by OpenAI that uses generative pre-trained transformer models to process natural language and produce human-quality text, code, analysis, and creative content. The platform supports text conversations, voice interactions, image analysis (Vision), image generation (DALL-E), code execution (Advanced Data Analysis), web browsing, file processing, and custom AI assistant creation (Custom GPTs). Available across web browsers, iOS, Android, macOS, and Windows. Free tier provides unlimited GPT-3.5 access. Paid tiers (Plus at $20/month, Team at $25/user/month, Enterprise at custom pricing) unlock GPT-4, faster speeds, and advanced features. Security: SOC 2 Type II, GDPR, CCPA, ISO 27001.
A complete inventory of ChatGPT capabilities across text, code, voice, vision, and data.
Text generation. ChatGPT writes emails, blog posts, reports, essays, social media content, product descriptions, and creative fiction. It adapts to specified tones, formats, and audiences. Feed it a company style guide and it matches your voice consistently. The prompt engineering guide explains how to structure requests for optimal output quality.
Code and development. The platform writes, debugs, and explains code in over 40 programming languages. Paste a stack trace and receive a corrected solution. Describe a function in plain English and get working, documented code. The Advanced Data Analysis feature executes Python in a sandboxed environment for immediate testing. Visit the API access page for programmatic integration.
Voice and vision. Voice Mode lets you speak naturally with ChatGPT using five distinct voice personalities. Vision processes uploaded images — photographs, charts, screenshots, handwritten notes, scanned documents. The voice mode and vision pages cover these features in depth.
Web browsing and data. ChatGPT searches the web for current information, processes uploaded CSV files and spreadsheets, generates data visualizations, and exports analysis results. It handles document summarization, translation between 95+ languages, and format conversion between file types.
Custom GPTs. The GPT Builder lets anyone create specialized AI assistants without coding. Upload reference documents, define behavior rules, and share your custom GPT with your team or the GPT Store. See the Custom GPTs page for builder documentation.
From free unlimited conversations to enterprise-grade AI deployment.
The free plan provides unlimited GPT-3.5 conversations, web browsing, access to the GPT Store, and limited GPT-4 access. No credit card required. This tier covers the needs of most casual users — students, hobbyists, and professionals exploring ChatGPT for the first time.
ChatGPT Plus at $20 per month unlocks GPT-4 with higher message limits, DALL-E image generation, Advanced Data Analysis, Advanced Voice Mode, and priority access during peak usage. Plus subscribers also get early access to new features before they roll out to free users.
Team plans at $25 per user per month add workspace collaboration, shared Custom GPTs, admin dashboard, and default data training opt-out. Enterprise pricing is custom and includes unlimited GPT-4, SSO via SAML, SCIM provisioning, dedicated support, and data isolation guarantees. The plan comparison page provides a feature-by-feature breakdown.
| Plan | Price | GPT-4 Access | Key Features | Best For |
|---|---|---|---|---|
| Free | $0 | Limited | GPT-3.5, web browsing, GPT Store | Casual users, students |
| Plus | $20/mo | Up to 80 msg/3hr | DALL-E, Voice, Data Analysis | Power users, professionals |
| Team | $25/user/mo | Higher limits | Workspace GPTs, admin tools | Small-medium teams |
| Enterprise | Custom | Unlimited | SSO, SCIM, data isolation | Organizations, regulated industries |
Available on every major platform with conversation sync across devices.
ChatGPT runs on web browsers (Chrome, Firefox, Safari, Edge), iOS (iPhone and iPad), Android phones and tablets, macOS (Apple Silicon and Intel), and Windows. Each platform offers the same core functionality — text chat, file uploads, voice input, image analysis, and model selection.
Sign in with the same account on every device and your conversation history syncs automatically. Start a conversation on your laptop during lunch, continue it from your phone on the commute home. Conversations, Custom Instructions, and memory persist across platforms.
For developers, the ChatGPT API provides REST endpoints accessible from any programming language. SDKs are available for Python, Node.js, and other popular frameworks. API access enables embedding ChatGPT capabilities into custom applications, internal tools, and automated workflows. Explore the integrations page for no-code connection options via Zapier and Microsoft 365.
The Federal Trade Commission provides guidelines on AI-assisted applications in consumer-facing contexts that businesses should review before deploying ChatGPT in customer interactions.
Enterprise-grade security infrastructure certified by independent auditors.
ChatGPT holds SOC 2 Type II certification, demonstrating verified security controls across data handling, access management, and system availability. Data in transit is encrypted with TLS 1.3. Data at rest is encrypted with AES-256. The platform complies with GDPR (European data protection), CCPA (California privacy), and ISO 27001 (information security management).
Users control their data through Settings > Data Controls. Disable chat history to prevent conversations from being used for model training. Conversations with history disabled are retained for 30 days for abuse monitoring, then permanently deleted. Enterprise and Team plans default to training opt-out and offer configurable retention policies.
Two-factor authentication protects account access. SSO via SAML and SCIM provisioning are available on Enterprise plans for centralized identity management. Read the security page for complete technical details and the privacy policy for data handling practices. Our AI safety page covers content filtering, RLHF alignment, and bias mitigation.
ChatGPT is an AI chatbot developed by OpenAI using generative pre-trained transformer technology. It processes natural language input and returns coherent, contextual responses across writing, coding, research, and analysis tasks. The free tier is available without a credit card and covers the needs of most everyday users.
The name "ChatGPT" combines two concepts: Chat (conversational interaction) and GPT (Generative Pre-trained Transformer, the neural network architecture underneath). When you type a message, ChatGPT tokenizes your input, passes it through dozens of transformer layers, and generates a probability-weighted response — all in one to five seconds. The model does not retrieve pre-written answers; it generates each response from scratch based on training and context.
What separates ChatGPT from older chatbots is generalization. Earlier systems required explicit rules for every scenario. ChatGPT learned from hundreds of billions of text tokens spanning virtually every domain of human knowledge — then was fine-tuned through reinforcement learning from human feedback (RLHF) to be helpful, harmless, and honest. That combination produces an assistant that handles novel questions it has never explicitly seen before.
The practical result: a single tool that replaces several specialized applications. Writers use ChatGPT where they previously used multiple grammar checkers, research tools, and outline generators. Developers use it where they previously opened Stack Overflow and documentation simultaneously. Students use it as an on-demand tutor available at 3 a.m. before an exam. The unifying thread is that ChatGPT meets users where they are, in plain language, without requiring technical fluency. See the use cases page for documented productivity examples, and the GPT technology page for architectural depth.
| Aspect | Detail |
|---|---|
| Developer | OpenAI, founded 2015 |
| Launch date | November 30, 2022 |
| Weekly active users | 200+ million |
| Underlying model | GPT-3.5 (free), GPT-4 / GPT-4o (paid) |
| Supported languages | 95+ |
A search engine retrieves links to existing pages. ChatGPT synthesizes information and produces original, context-aware text. You can have a back-and-forth dialogue, ask follow-up questions, and request edits or different formats — none of which search engines support natively. ChatGPT also handles tasks like writing, coding, and analysis that search fundamentally cannot perform.
No. ChatGPT's responses are generated probabilistically, meaning slight variation is built in. The same prompt asked twice may produce different wording, emphasis, or structure. Context within a conversation also changes responses — earlier messages shape what the model generates next. Custom Instructions let you personalize default behavior so responses consistently match your preferences and workflow.
A chat with ChatGPT begins the moment you type in the message box and press Enter. The model reads your full conversation history — every prior message in the session — before generating each reply. Conversations can run for dozens of exchanges, branching across topics, until you start a new chat or the context window fills.
The mechanics of chatting with ChatGPT are deliberately simple by design, but the depth available to experienced users is substantial. At its most basic: open a new chat, type your message, read the response, and reply. The model treats the entire thread as a single document, attending to all prior messages when composing each reply. This means you can refer to earlier statements ("the code you wrote above") without re-explaining context.
Conversation quality scales with prompt quality. Vague questions produce generic answers. Specific, structured prompts with clear context, constraints, and desired format produce precise, useful output. The difference between "summarize this" and "summarize this 500-word article in three bullet points for a non-technical audience" is the difference between a mediocre result and a publish-ready one. The prompt engineering guide covers dozens of techniques for structuring high-value prompts.
ChatGPT supports multi-turn dialogue across modalities. Within a single chat session you can switch between typing, uploading a file, sharing an image, and using voice — the model maintains context across all of them. You can paste a screenshot of a chart, ask what it shows, then ask for follow-up analysis, then request a Python script to replicate the chart from raw data — all without leaving the conversation. For extended projects, the Custom GPTs feature lets you configure persistent behavior so every chat session starts with the same baseline instructions without manual setup.
| Aspect | Detail |
|---|---|
| Context window (GPT-3.5) | ~4,096 tokens (~3,000 words) |
| Context window (GPT-4) | Up to 128,000 tokens (~96,000 words) |
| Input types | Text, images, files, voice, URLs |
| Session persistence | Saved to history, synced across devices |
| Max turns per chat | No hard limit (context window governs) |
By default, ChatGPT does not carry memory between separate chat sessions — each new chat starts fresh. However, the Memory feature (available on Plus and Team plans) lets ChatGPT retain facts across sessions: your profession, preferences, recurring project details. You can view, edit, and delete memory entries at any time from Settings. Custom Instructions provide a lighter-weight alternative that applies a standing context to every chat without storing conversational details.
Once a conversation fills the context window, the oldest messages are dropped from the model's active attention. ChatGPT will not explicitly warn you — it will simply stop referencing early parts of the conversation accurately. The solution is to start a new chat, paste a summary of the key points from the old session, and continue. For very long projects, using GPT-4's 128,000-token context window significantly extends the point at which this becomes an issue.
GPT is the neural network model; ChatGPT is the application built on top of it. GPT generates text given an input sequence. ChatGPT adds the conversation interface, safety alignment, memory, browsing tools, image generation, and voice capabilities that turn a raw language model into a practical assistant. You interact with ChatGPT; ChatGPT relies on GPT.
The distinction between ChatGPT and GPT matters because it clarifies what the model can and cannot do on its own. A base GPT model — without fine-tuning or alignment — will complete text plausibly but unpredictably. It will continue a story, finish a sentence, or answer a factual question, but it has no concept of helpfulness, safety, or conversational turn-taking. It is a sophisticated autocomplete engine.
What OpenAI did with ChatGPT was layer three additional training phases on top of the base GPT model. Supervised fine-tuning trained the model on example conversations between humans and ideal AI assistants. Reinforcement learning from human feedback (RLHF) used human raters to score model outputs and trained the model to prefer higher-rated responses. The result is a model that follows instructions, refuses harmful requests, maintains a helpful tone, and produces responses calibrated for dialogue rather than text completion. The AI safety page explains the RLHF alignment process in detail.
In practical terms, the ChatGPT-GPT relationship means that the quality of your experience depends on both layers. GPT-4 versus GPT-3.5 is a model-level difference — GPT-4 has more parameters, better reasoning, and a larger context window. But the ChatGPT layer — browsing, DALL-E, data analysis, custom instructions, memory — operates above the model. Some ChatGPT features run on GPT-4 only; others are available on both. The GPT models page maps which capabilities are tied to which model version, and the plan comparison shows which plans unlock each feature.
| Aspect | Detail |
|---|---|
| GPT role | Neural network that generates text |
| ChatGPT role | Product interface with tools, safety, and memory |
| Fine-tuning method | RLHF (Reinforcement Learning from Human Feedback) |
| Current models in ChatGPT | GPT-3.5, GPT-4, GPT-4o |
| API access to GPT | Available separately via OpenAI API |
Yes. Developers can access GPT models directly through the OpenAI API, bypassing the ChatGPT interface entirely. API access allows you to set your own system prompts, control model parameters (temperature, top-p, max tokens), and integrate GPT into custom applications. The API access page covers authentication, endpoint documentation, and rate limits. API usage is billed per token rather than as a flat subscription.
Free plan users default to GPT-3.5, with a limited daily allocation of GPT-4 messages. ChatGPT Plus, Team, and Enterprise subscribers can select their model from the conversation interface — GPT-3.5, GPT-4, or GPT-4o. GPT-4o is OpenAI's fastest multimodal model, optimized for lower latency. The model selector appears at the top of the ChatGPT conversation window for users with access to multiple models.
Join 200 million weekly users. Create a free account in under two minutes and start your first AI conversation.
Get Started FreeQuick answers to the most common ChatGPT questions.
ChatGPT is an AI chatbot built on generative pre-trained transformer (GPT) models developed by OpenAI. It processes natural language input and produces human-quality text, code, analysis, and creative content. The platform supports text conversations, voice interactions, image processing, code execution, web browsing, and custom AI assistant creation. Over 200 million people use ChatGPT weekly across web, mobile, and desktop applications.
ChatGPT writes content (emails, blogs, reports, creative fiction), generates and debugs code in 40+ languages, translates between 95+ languages, analyzes images and documents, processes data with Python execution, creates images via DALL-E, holds voice conversations, browses the web, and builds custom AI assistants. Visit the use cases page for detailed examples across industries.
ChatGPT offers a free tier with unlimited GPT-3.5 conversations. ChatGPT Plus costs $20/month with GPT-4, DALL-E, and advanced features. Team plans are $25/user/month with workspace tools. Enterprise pricing is custom with unlimited GPT-4, SSO, and dedicated support. The plan comparison provides a complete feature breakdown.
Yes. ChatGPT has official apps for iOS (iPhone and iPad) and Android devices. Desktop apps are available for macOS and Windows. All platforms offer the same features — text chat, file uploads, voice input, image analysis, and model selection. Conversation history syncs automatically across all devices when signed in to the same account.
ChatGPT uses TLS 1.3 encryption in transit and AES-256 at rest. The platform holds SOC 2 Type II certification and complies with GDPR, CCPA, and ISO 27001 standards. Users can disable chat history in Settings to prevent training use. Enterprise plans add data isolation, SSO, and configurable retention. Read the privacy policy and security page for complete details.
Visit the ChatGPT platform and click Sign Up. Create a free account using your email or a Google, Microsoft, or Apple login — no credit card required. Once registered, click New Chat and type your first message. ChatGPT responds within seconds. The getting started guide walks through account creation, interface navigation, and first-conversation tips in detail.
GPT is the underlying language model — the neural network trained on billions of text samples. ChatGPT is the consumer product built on GPT: it adds a conversation interface, safety filtering, memory, browsing, DALL-E image generation, and voice capabilities. GPT provides the intelligence; ChatGPT makes it accessible and safe for everyday use. Visit the GPT technology page for architectural details.
Yes, with limitations. ChatGPT allows some anonymous access, but conversation history does not save, Custom Instructions cannot be configured, and features like file uploads and Custom GPTs are restricted. Creating a free account unlocks the complete free tier: persistent history, cross-device sync, GPT Store access, and limited GPT-4 messages. Registration takes under two minutes with social login.
Deep-dive into specific aspects of the ChatGPT platform.
ChatGPT