ChatGPT
General-purpose AI assistant for chat, research, file analysis and code.
What it is
A horizontal assistant that covers the widest range of everyday tasks. It is not the cheapest option at the top tier, but it is the one most teams can adopt without changing how they already work. Choose it when you want one subscription to cover writing, data, images and code rather than five specialised tools.
Pricing updated 2026-11-10 Official pricing
ChatGPT pricing explained
The official page splits individual plans into four steps, and they differ on capacity rather than on core intelligence: each step up buys more messages, larger memory and storage, and access to the heavier models. The free tier is enough to judge whether the tool fits your work, but image generation, memory and upload limits make it uncomfortable as a daily driver. Plus is the step most individuals should start with. The top individual tier only pays off if the assistant is in use for most of the working day and long reasoning runs are routine. Business and enterprise pricing sits behind a separate tab and is quoted per seat. One caveat: the plan cards above are the snapshot taken on the update date, and the official page has since renamed its individual tiers, so confirm names and prices on the official page before buying.
Free-tier limits are taken from the official pricing page on the update date above. They change often — confirm before you commit to an annual plan.
ChatGPT key features
Multi-turn conversation
Keeps the thread in context, so a long refinement session replaces dozens of separate prompts. This is the part most people actually pay for.
Web search with citations
Answers questions past the training cut-off and links the pages it used, which makes the output checkable instead of merely plausible.
File and image understanding
Reads PDFs, spreadsheets and screenshots, then answers questions about them — useful when the material is a document rather than a question.
Image generation and editing
Produces or edits images inside the same conversation, so a campaign concept and its visual can be iterated in one place.
Voice conversation
Two-way spoken dialogue on mobile and desktop. Practical for hands-free drafting and for practising a language out loud.
Custom GPTs and Projects
Turns a repeated workflow into a reusable assistant with fixed instructions and files, which is how a team gets consistent output instead of twelve slightly different answers.
ChatGPT: how it works
Transformer decoder
Generates one token at a time, attending over the whole context window. That architecture is why quality degrades gradually as a conversation gets very long rather than failing at a hard limit.
Pretraining, then alignment
Language and world knowledge come from large-scale text pretraining; tone, refusal behaviour and instruction following come from later human-feedback tuning. The two stages explain why a model can be knowledgeable and still evasive.
Tool use instead of recall
For fresh facts, arithmetic and files, the assistant can call search, code execution and retrieval rather than answering from memory. That is the practical difference between a chatbot and an assistant.
Context and memory
The visible conversation, uploaded files and saved memory are all passed back into the context window. Memory makes answers feel personal; it also means stale saved facts can quietly bias later replies, which is why the memory list is editable.
ChatGPT core strengths
Breadth in one subscription
Writing, research, files, images, voice and code sit behind the same login. For a small team, that replaces three or four separate tools and the subscriptions that come with them.
An ecosystem, not just a chat box
Custom GPTs, a documented API and first-party desktop and mobile apps mean the same product can be a personal tool and a building block inside someone else workflow.
Multimodal in one thread
Image, document and voice input land in the same conversation as text, so a single thread can move from a screenshot to a written summary without exporting anything.
Mature and predictable
Rate limits, tiers and behaviour are documented and stable enough to plan around. Boring reliability is an advantage once a tool is part of a daily process.
ChatGPT compared with alternatives
| Tool | ChatGPT | Claude | Gemini | DeepSeek |
|---|---|---|---|---|
| Best for | One assistant for mixed everyday work — writing, research, files, images | Long documents and writing that has to hold up across many turns | People already living in Google Workspace, Search and Android | Reasoning-heavy work on a budget, or self-hosting when data cannot leave |
| Free tier | Usable but capped: newest model limited, few uploads per day | Daily message cap with limited uploads | The most generous of the group for everyday use | Free chat with no paid tier required |
| Inputs and outputs | Text, image, voice in; text, image, voice out | Text and image in; text out — no image generation | Text, image, audio and video in; image generation out | Text in and out — no image generation |
| Ecosystem and API | Custom GPTs, documented API, first-party desktop and mobile apps | Projects, Artifacts, API; smaller plugin ecosystem | Gems, Workspace add-ons and API — strongest inside Google apps | Open weights plus a low-cost API |
| Paid entry price | $20 / month | $20 / month | $19.99 / month | Free; API billed per token |
Written from the official product pages listed under Sources, not copied from another directory. Prices and limits reflect the update date above; features ship faster than reviews, so treat the list as a starting point and confirm on the official site.
Pros and cons
Strengths
- Widest feature surface of any assistant
- Mature ecosystem: GPTs, API, desktop and mobile apps
- Strong multimodal output — text, image, voice in one place
Limitations
- Usage caps differ per tier and change often
- Advanced reasoning sits behind the Pro tier
- No EU data residency on consumer plans
ChatGPT use cases
Writing and rewriting
Drafting, tightening and changing the register of a text — the most common use, and the one where the long context matters most.
Research and summarising
Reading several long sources and producing a comparison with links, instead of skimming each one and hoping nothing important was missed.
Code and debugging
Explaining unfamiliar code, writing tests and narrowing a failure to a specific line, in the same thread where the error was pasted.
Data analysis
Running small analyses and producing charts from an uploaded spreadsheet without writing the plumbing yourself.
Internal knowledge Q&A
A Custom GPT or Project loaded with a team policy set answers the same question the same way, which is what makes it usable as an internal reference.
Learning and practice
Explaining a topic at the level the learner asks for, then testing them on it — the loop is faster than searching for a good tutorial.
