I Cancelled ChatGPT Plus, Claude Pro and Gemini Advanced. Here's What I Use Instead.
For about eight months I paid for all three: ChatGPT Plus for daily work, Claude Pro for anything long or code-heavy, and Gemini Advanced because it kept showing up in Google Workspace and I figured I might as well use what I was already half-paying for. That's $60/month, three logins, and a constant low-grade decision fatigue about which tab to open for which task.
Why I kept all three instead of picking one
The honest reason: each one was genuinely better at something. ChatGPT was my default for quick questions and general coding. Claude was clearly better the moment a task involved a long document or a codebase I needed it to actually track carefully across a long conversation without losing the thread. Gemini won on anything tied to Docs or Sheets, and on tasks where I wanted it to pull in current web results.
None of them was "the best model" — they were each the best model for something, and canceling any one of them meant losing that something.
What finally made me switch
Two things. First, the arithmetic: $60/month is $720/year to keep three tools "warm" for tasks I only hit occasionally with any one of them. Second, and more annoying: I kept forgetting which subscription I needed before starting a task, so I'd start in ChatGPT, hit a wall on a long document, and have to re-explain the whole context to Claude in a new tab. The switching cost wasn't just money, it was re-explaining myself three times a week.
What I use now
I moved everything to AIWITH.CHAT, which puts GPT-5.6, Claude, Gemini, DeepSeek and Kimi behind one login with model switching built into the same conversation — no re-pasting context, no juggling three billing pages. Plans start under $10/month, meaning I'm now paying less than one of my old three subscriptions to get access to all of them plus a couple I hadn't tried before (DeepSeek turned out to be genuinely good for the cheap, high-volume stuff I used to default to GPT for out of habit).
What actually changed day to day
- I stopped pre-deciding which tool to open. I start typing and switch models mid-conversation if the first one isn't landing.
- I stopped losing context between tools — switching models doesn't mean starting a new thread.
- My monthly AI spend dropped from $60 to under $10, and I use more models than I did before, not fewer.
The three-subscription era wasn't wrong, exactly — each tool really was good at what it was good at. It was just an expensive way to buy optionality I could get for a fraction of the price in one place. If you're paying for more than one AI subscription right now, it's worth checking whether you actually need three logins or just one that doesn't box you into a single lab.