Module 01
8 exercises
Basics
Get productive with ChatGPT fast — the core mental model and first prompts.
2 lessons
Start module
OpenAI · Course
Learn ChatGPT.
ChatGPT is OpenAI's assistant, strong at broad general tasks, image and voice, and a deep tool ecosystem (Memory, Projects, Canvas, Deep Research). This hub breaks ChatGPT down by intent — pick what you're doing, then drill into tested knowledge cards.
Course syllabus
A practical path through ChatGPT.
Move through the modules in order, or return to the part of your work that needs attention now.
Module 02
12 exercises
Coding
Generate, debug, and ship code with ChatGPT — from snippets to full features.
3 lessons
Start module
Module 03
8 exercises
Writing
Draft, edit, and structure writing with a consistent voice.
2 lessons
Start module
Module 04
4 exercises
Research
Synthesise sources and reason through complex questions.
1 lesson
Start module
Module 05
4 exercises
Workflows
Multi-step systems that chain prompts into a repeatable outcome.
1 lesson
Start module
Module 06
16 exercises
Hidden Features
Capabilities most people miss — Memory, Projects, Canvas, Deep Research.
4 lessons
Start module
Module 07
4 exercises
Advanced
Context control, tool use, and reasoning models for power users.
1 lesson
Start module
Module 08
4 exercises
Mistakes
The common errors that make ChatGPT output worse — and how to avoid them.
1 lesson
Start module
How to use ChatGPT well
What ChatGPT is actually good at
ChatGPT is the most general-purpose of the mainstream assistants: it drafts and edits writing, explains hard concepts, reasons through problems, generates and runs code, and works with images, voice, and uploaded files in one place. That breadth is also the trap — because it will attempt almost anything, people accept a confident answer to the wrong question. The way to use ChatGPT well is to stop treating it as a search box and start treating it as a capable collaborator that needs context, a clear goal, and a defined output format.
The single biggest lever is picking the right model for the job. The fast default is ideal for everyday drafting, summarising, and quick lookups. For multi-step logic, tricky maths, or careful planning, switch to a reasoning model and let it think before it answers. Reaching for a heavyweight reasoning model to reformat a paragraph wastes time; using the fast model for a subtle proof gets you a fluent, wrong answer.
A simple structure for better prompts
Most disappointing ChatGPT answers come from one habit: a one-line request with no context. A reliable prompt has three parts — the context (who you are, what you are working on, any constraints), the goal (the specific outcome you want), and the format (length, structure, tone, audience). 'Write about onboarding' produces filler. 'You are helping a B2B SaaS team. Write a 6-step onboarding email sequence for trial users who signed up but never logged in — each email under 90 words, one clear CTA' produces something you can ship.
Two moves compound this. Set your custom instructions once so ChatGPT already knows your role, preferences, and how you like answers formatted on every new chat — you stop repeating yourself. And when a task is genuinely complex, resist stuffing everything into one mega-prompt; break it into steps and let each build on the last.
The features most people never touch
The gap between casual and power users is almost entirely the hidden features. Memory lets ChatGPT carry facts across conversations — useful when curated, noisy when it remembers the wrong things, so review it periodically. Projects group related chats and files under one shared context, which keeps a long piece of work coherent instead of scattered across a dozen threads. Canvas pulls writing and code into a side-by-side editor so you revise in place rather than regenerating the whole answer in chat. And Deep Research runs long, cited investigations across many sources — powerful, but only if you scope the question tightly before you set it running.
Most 'ChatGPT isn't that useful' complaints come from someone who has only ever used the plain chat box. Learning these four features is the fastest upgrade available.
Turn one-off prompts into repeatable systems
Once a prompt earns its keep — a weekly report, a support-reply drafter, a code-review checklist — stop retyping it. A custom GPT packages the instructions, tone, and any reference files into a reusable assistant you (or your team) can open and run in one click. That is the jump from using ChatGPT as a toy to running it as part of a workflow. Pick the intent below that matches what you are trying to do and drill into the tested knowledge cards underneath.
Frequently asked
Is ChatGPT free to use?
Yes. OpenAI offers a free tier with access to a capable default model, and paid plans (Plus, Team, Pro) that add higher limits, the strongest reasoning models, and features like advanced Deep Research. For most individual learning and everyday tasks the free tier is enough to get productive; the paid tiers pay off when you rely on the reasoning models or hit message limits.
Which ChatGPT model should I use?
Use the fast default model for everyday writing, summarising, and quick questions — it is quick and light on your limits. Switch to a reasoning model for anything with multiple steps, careful logic, maths, or planning, where being right matters more than being fast. Matching the model to the task is the highest-leverage habit in ChatGPT.
How do I write a good ChatGPT prompt?
Give it context, a clear goal, and the format you want back. State who you are and what you are working on, say exactly what outcome you need, and specify length, structure, and tone. Add one example of what 'good' looks like when you can, and avoid piling several unrelated requests into a single mega-prompt — break complex work into steps instead.
Can ChatGPT write and run code?
Yes. ChatGPT generates code across most languages and, with its data-analysis tool, can execute Python in a sandbox to analyse files, run calculations, and produce charts. For debugging, paste the exact error and the smallest snippet that reproduces it rather than describing the symptom — that single change dramatically improves the fixes you get back.
Does ChatGPT remember previous conversations?
Only if Memory is on. With Memory enabled, ChatGPT can carry selected facts and preferences across chats; without it, each new conversation starts blank apart from your custom instructions. Memory is worth curating — keep what is genuinely reusable and clear out anything stale or incorrect so it does not skew future answers.
How is ChatGPT different from Claude and Gemini?
ChatGPT is the broadest generalist with the deepest feature ecosystem (Memory, Projects, Canvas, Deep Research) and strong multimodal support. Claude tends to lead on long-form writing, coding discipline, and careful reasoning; Gemini stands out for very long context and tight Google Workspace integration. Many people use more than one and route each task to the tool that fits it best.
Course checkpoint
ChatGPT Master Quiz
Ten questions spanning everything ChatGPT — prompting, models, coding, writing, research, custom GPTs, and the mistakes that cap most people's results. Take it as a diagnostic: whatever you miss tells you which topic to read next. No sign-up, instant score, and you can share how you did.