ChatGPT Plus vs Claude Pro vs Gemini Advanced: which $20 plan

The honest starting point

These three plans all sit at roughly $20 a month, all give you a frontier model, and all are good enough that you would get useful work done with any of them. Anyone who tells you one is objectively best across the board is either selling something or has only used one of them properly.

A note on names before we start. Gemini Advanced was folded into Google AI Pro, so if you are searching for Gemini Advanced you are looking for that plan. Claude also sells much heavier Max tiers, and ChatGPT has a Pro tier around ten times the price. This comparison is about the $20 rung, since that is where almost every real decision happens.

The tiebreakers that actually decide it are not benchmark scores. They are, in rough order of how much they matter day to day:

1. Which ecosystem you already live in. Gmail and Docs all day, or Microsoft, or neither.
2. What is bundled. One of these three includes 2TB of cloud storage, which is not a small detail.
3. Whose writing voice needs less editing. Completely subjective and completely decisive if you write for a living.
4. Whether you code, and where. In a terminal, in an editor, or in a chat window.

Verify current pricing on each vendor's site before you buy. Plan names and inclusions in this category change several times a year, and the bundled extras change more often than the prices do.

What each plan actually includes

Feature lists blur together, so this table focuses on what is different rather than what all three share. Every one of them gives you a frontier model, file uploads, image understanding, web access, and a mobile app.

	ChatGPT Plus	Claude Pro	Google AI Pro
Around	$20 / month	$20 / month, cheaper annually	$20 / month
Best known for	Breadth and ecosystem	Writing quality and code	Long context and Google apps
Signature feature	Custom GPTs, code execution, connectors	Artifacts, Projects, Claude Code	Gemini inside Gmail, Docs, Drive, Sheets
Context window	Large	Large, with much larger tiers available	The largest of the three by a wide margin
Bundled extras	Image and video generation allowances	Coding agent access with limits	2TB cloud storage, video generation, NotebookLM limits
Weakest at	Sourced research with citations	Ecosystem breadth and integrations	Consistency of tone in long writing
Free tier	Good	Good	Good

Look at the bundled extras row again. If you already pay Google for 2TB of storage, the AI plan effectively costs you the difference rather than the full price. That single fact settles the decision for a lot of people before any model comparison starts, and almost no comparison article mentions it.

Where each one genuinely wins

ChatGPT Plus wins on breadth

It is the tool that does the most different things acceptably well. Code execution on your uploaded data, image generation, voice, connectors into workplace tools, custom GPTs for repeatable jobs, and by far the largest body of community knowledge when something goes wrong. If you want one assistant that will handle whatever lands on your desk, this is the safe pick, and "safe" is a real virtue when the alternative is being stuck.

It is also the best pick for anyone building workflows or automations, because the API and the tooling around it are the most mature and the most documented.

Claude Pro wins on writing and reading

Two things stand out consistently. First, the default prose needs less editing. It holds a register, it hedges less mechanically, and it produces first drafts that read like a competent human wrote them rather than like a template got filled in. If you write for other people to read, test this before you commit to anything else.

Second, it is very good at reading. Long documents, dense contracts, code review, sprawling research notes: Claude handles synthesis carefully and is more willing to say what a document does not say. For developers, Claude Code brings that same quality into the terminal, and it is included with the plan at usage limits that suit a solo developer.

The catch is breadth. Fewer integrations, a smaller ecosystem, and tighter usage limits than people expect from a $20 plan if you push it hard.

Google AI Pro wins on scale and on Google

The context window is the headline. Gemini can take in far more material at once than the other two, which changes what is possible: a whole codebase, a year of meeting notes, a 900-page manual. Long-context work is where it feels like a different class of tool rather than a slightly different one. Our context window explainer covers why bigger is not automatically better, but for the "read all of this and find the pattern" job, it is.

Then there is the integration. Gemini in the side panel of Gmail, Docs, Sheets, and Drive means the assistant is where the work already is, with access to your files without uploading anything. Add 2TB of storage, video generation, and higher NotebookLM limits, and the value per dollar is the strongest of the three for anyone in the Google ecosystem.

The weakness is consistency. Output quality varies more between tasks than the other two, and long-form writing needs more editing.

Where each one genuinely loses

This is the part that matters more than the strengths, because you will feel the weakness of whichever plan you pick every single week.

Plan	The thing that will annoy you
ChatGPT Plus	Writing that needs a pass to remove its default rhythm, and research answers that summarise rather than cite
Claude Pro	Hitting usage limits sooner than you expect on heavy days, and a thinner ecosystem when you want an integration
Google AI Pro	Uneven output between tasks, and a product surface that changes names and layouts often

None of these are dealbreakers. They are the friction you are choosing. It is easier to accept friction you picked deliberately than friction you discover in month three.

Pick one in sixty seconds

Answer the first question that applies to you and stop reading.

• Do you live in Gmail and Google Docs all day? Google AI Pro. The integration and the storage bundle make it the best value, and you will use it more because it is already there.

• Do you write things other people read, professionally? Claude Pro. The editing time you save is worth more than any feature list difference.

• Do you build software, mostly in a terminal or reviewing code? Claude Pro, for Claude Code and code review quality. If you work primarily inside an IDE, note that a dedicated in-editor assistant may serve you better than any of these three, and read our AI for coding guide.

• Do you need one tool that can attempt anything, including data analysis, images, and automations? ChatGPT Plus.

• Do you work with very long documents, video, or huge codebases? Google AI Pro, for the context window.

• Is most of your work research where every claim needs a source? None of these three are ideal. Look at a citation-first tool and read Perplexity vs ChatGPT first.

• Do you genuinely need two of them? Read the next section before you pay for two.

The two-subscription problem

Here is what actually happens to people who take this decision seriously. They pick one, use it for a few months, and hit its weakness. Then they add a second plan to cover the gap. Now they are at $40 a month, with their history split across two products, pasting prompts between tabs to compare answers, and never quite sure which tool to open first.

The instinct is right. Different models really are better at different jobs, and the ranking changes with every release, so committing to one vendor means accepting whatever they ship next. The instinct is also expensive, and it gets worse each time a new model makes headlines.

The alternative worth knowing about: a multi-model workspace gives you these models behind one subscription. Whizi includes GPT, Claude, Gemini, Grok, and DeepSeek in one place, so you can start a draft in Claude, ask GPT to poke holes in it, and hand the 300-page appendix to Gemini, in one conversation and on one bill. How to use multiple AI models together covers the workflow, and Whizi vs ChatGPT Plus, Whizi vs Claude Pro, and Whizi vs Gemini Advanced give the direct comparisons.

The honest counterpoint: a single vendor plan gives you that vendor's deepest product features, and some of them are excellent and not replicable elsewhere. Claude Code, Gemini inside Google Docs, and ChatGPT's connectors are product features, not just model access. If one of those is central to your day, buy that plan and do not overthink it.

The decision rule is simple. If one product's features are the reason you want it, buy that product. If model access is the reason, buy model access. Most people who own two subscriptions are in the second group and bought the first thing.

Test all three before you commit

Every one of these has a free tier that is good enough to judge the model, even if the limits are tighter. Spend one afternoon rather than three months of subscription fees.

Take four prompts from real work: a piece of writing you will actually send, a document you need summarised accurately, a problem where you already know the correct answer, and whatever your most repeated task is. Run all four through all three free tiers, on the same day, with identical prompts.

Score only three things per answer: how many minutes to a usable result, how many edits it needed, and whether you would have caught a mistake if you had not been checking. That last one is the most revealing and the one people never measure.

Then total your existing AI spend before adding anything with the savings calculator, and pick a plan from the pricing page once the numbers and the test agree with each other.