People ask me this one a lot. And my honest answer is: for most everyday things, they're more similar than different. But there are a few things worth knowing.
ChatGPT is made by OpenAI. It's the one that went viral in late 2022 and introduced most people to what AI assistants could actually do. It's the one your dad has probably heard of.
Claude is made by Anthropic. Anthropic was founded by some former OpenAI researchers who wanted to build AI differently — with more focus on safety and predictability. Whether that fully works is a whole other conversation. But that's the intent.
Gemini is made by Google. It's the third big name you'll keep hearing, and if you're already deep in Google-land — Gmail, Docs, Android — it's the one built by the company you already use. It deserves a seat at this table, so I've stopped pretending this is a two-horse race.
ChatGPT
Made by OpenAI
First to go mainstream (2022)
Built-in image generation (ChatGPT Images 2.0 — the DALL-E era is over)
Current top model: GPT-5.5
Free tier available
Claude
Made by Anthropic
Generally better at long documents
No image generation (charts and diagrams yes, photos no)
Current top model: Claude Opus 4.8
Free tier available
Gemini
Made by Google
Built into the Google world — a Google account is all you need
Built-in image generation (the image model is called Nano Banana — yes, really)
Current top model: Gemini 3.1 Pro
Free tier available
In practice, all three can write, explain, summarise, code, analyse, and generally be useful. All three have free versions that are genuinely good. All three make mistakes. And all three are getting better fast enough that anything I write here will be slightly outdated by next month. For the record, as I update this in July 2026 the current top models are OpenAI's GPT-5.5, Anthropic's Claude Opus 4.8 and Google's Gemini 3 — but the version numbers will have moved on by the time you read this, and honestly, it doesn't matter.
I use Claude. Not because I think it's definitively better — but because it's what I started with, and because Anthropic's approach to building AI carefully resonates with me. That's honestly it. Personal preference, not objective truth.
If you're just starting out, try two of them. Spend a week with each on the same kinds of tasks. See which one's responses feel more useful to you. The answer will probably be different for different people.
The thing I'd really rather you focus on: how you talk to these tools matters enormously. A vague question gets a vague answer. A specific, well-framed question gets a genuinely useful one. That skill — sometimes called "prompting" — is worth more than any particular tool choice.
Facts checked July 2026.