Three chatbots. Same price of $20 for their subscriptions. The convergence is almost funny considering how different the offerings are.
The same price does not mean the same product. I paid for all three and ran the same work through each. They are not interchangeable as you’ll soon find out.
Pick wrong and you’ll spend months wondering why everyone else seems to be getting more out of this. Pick right and the 20 dollars pays for itself by Tuesday. So here’s what actually separates them.
The whole picture, at a glance
My scores after months of use:
Rule two out first
Do not ask which is best. Ask what you cannot live without. A few need to settle it instantly:
- Need video? Only Google does it, and does it well. Claude and ChatGPT do not. Decision made.
- Need images? Claude is out entirely. ChatGPT makes the best ones; Google is a close second.
Two needs knock out two plans before the real contest begins.
The fight over words
For most people, the daily job is writing. If that’s your work, the choice isn’t close. Claude writes best of the three models, by a mile. Here are its standout features:
- Sounds human, not assembled: The others read templates. Claude has a voice. Its default mode of economical word usage and precise intent delivery puts it in a league of its own.
- Learns your voice: Feed it your old writing samples and watch it create every draft that comes back sounding like you. This is exclusive to Claude.
- Holds tone across long pieces: Consistent from first line to last, due to having long contexts in the chat interface.
Caveat: Best writer, not best editor. You can’t hand-edit its drafts, only ask for changes.
If the words are the product, choose Claude.
The fight over code
Closest race of the bunch. Two real winners here, not one.
Claude and ChatGPT are both legit. Either ships code you can actually use, tracks what’s happening across files, and holds up when you let it run loose.
Gemini’s the odd one out: it gets shaky on big, messy, unfamiliar codebases. The huge context window (1M+) is handy for reading a whole repo, but doesn’t save it when the problem’s actually hard. Fine for a script, not for the gnarly stuff.
Between the two that matter:
- Cleaner code: Claude. Thinks before it types, giving a tighter structure. The one most people reach for when it’s serious.
- Better workflow: ChatGPT. Codex lives inside your setup, so code lands where you already work. Claude Code is close, but ChatGPT edges it here. Gemini CLI is just a placeholder compared to these two.
Catch with Claude: usage limits dry up fast on long sessions. Plan around it.
Read more: 23 tips to save Claude Code tokens
The fight over research
Every model says it can research. They don’t mean the same thing by it. Ask all three the same question and the split shows up right away.
- Gemini casts the widest net. Google’s search DNA is doing the work, so it reaches further and comes back with the freshest stuff. If you need what happened this week, it’s the one that actually knows.
- Claude is the one that makes the pile make sense. It doesn’t just hand you a stack of links. It reasons over them, connects them, and tells you what they add up to. Linguistically the outputs are above all else.
- ChatGPT is the steady all-rounder. Good citations, thorough, web access that works. It’s rarely the sharpest at either finding or reasoning, but it doesn’t faceplant either. One catch though: when it’s wrong, it’s wrong with total confidence, the same calm voice it uses for facts. Verify it
The bigger catch: casting a wide net and knowing what’s in it aren’t the same skill. Gemini finds the most. It doesn’t always know what matters.
What it remembers
A model that forgets you every session makes you do the same setup forever. Memory is the difference between a tool and an assistant.
- Most mature: ChatGPT. Builds a real profile over time, carries it across chats, and you feel it.
- Most personal: Claude. Memory plus the voice-learning skill means it doesn’t just remember facts, it remembers how you sound.
- Most contextual: Gemini. Less standalone memory, but it leans on everything Google already knows about you.
Caveat: the same memory that helps can drift. All three occasionally remember the wrong thing confidently.
What each plugs into
Connections allow these models to further enhance their capabilities. While some of the models offer limited (albeit robust) integration, the others are a lot more liberal when it comes to connections:
- Claude: connects to almost everything, because it includes Zapier.
- ChatGPT: supports several types of integrations barring Zapier.
- Gemini: The entire Google Ecosystem. Gmail, Drive, YouTube and any other service/product that falls under the Google umbrella can be connected with Gemini.
The overlooked feature
Two plans hand you a desktop agent that works on your real files. This changes everything.
- Claude Cowork points at a folder and reads, edits, and builds files. Best-in-class for knowledge work.
- ChatGPT Codex is just as capable, a touch less tidy. Either makes you happy.
- Gemini Spark is cloud-only and locked to higher tiers. Not in this plan yet.
Google is a bundle
Google AI Pro is more like Pandora’s box than an LLM subscription: video tools (Veo3 & Omni), music, NotebookLM, YouTube Premium, a health coach, cloud credits, 5TB Google Drive storage and much more.
Several are $10 products alone. The standalone value clears the price, if you use the pieces.
But in terms of the value it offers as an LLM, it is pretty lacklustre.
The Ceiling
Features mean nothing if you hit a wall, and you will hit one. The real question is how soon, and that single fact can decide your whole subscription.
The biggest divide is what happens when you hit it:
- ChatGPT and Gemini quietly downgrade you to a weaker model and let you keep going.
- Claude just stops. No fallback, no downgrade, you wait for the reset.
The takeaway: If you work in long, intense bursts, Claude’s wall arrives first and hits hardest. ChatGPT is the most predictable, while Gemini is the most forgiving.
Here’s how the limits actually work:
The table now carries the per-plan detail, and the prose just makes the one point that matters.
Which $20 AI Plan Should You Buy?
Match the plan to how you work.
- Claude: best writing, the Cowork agent, near-universal integrations. It runs out first.
- ChatGPT: best images, steadiest limits, the safe all-rounder.
- Google: video, a real editor, the deepest bundle, if you already live in Google.
Here’s the full comprehensive comparison between the subscriptions:
I specialize in reviewing and refining AI-driven research, technical documentation, and content related to emerging AI technologies. My experience spans AI model training, data analysis, and information retrieval, allowing me to craft content that is both technically accurate and accessible.
Login to continue reading and enjoy expert-curated content.
