Why I Switched from ChatGPT to Claude and Gemini

I’ve been using ChatGPT since shortly after it was released back in 2022. As a writer, I was initially deeply unimpressed with the output. But over time, the models got better. I found myself integrating ChatGPT into my workflows, especially to repurpose content (such as turning a blog post into a social post, which I then edit).

But over the past month, I've switched from ChatGPT to both Claude and Gemini. My ChatGPT subscription will renew in about two weeks, and I'll probably let it expire. At most, I'll hang onto it for one additional month to make sure everything is set up with my new systems.

Here's why I decided to make the switch.

Other models have become more competitive

For a long time, ChatGPT was the dominant model. It caught my attention when Sam Altman, CEO of OpenAI called a "code red" in early December. Code red meant "all hands on deck, we need to make ChatGPT better."

This was in response to Google's latest release of Gemini 3. Anthropic released its latest model, Opus 4.5, around the same time.

I'm a regular listener of the podcast Hard Fork, and they covered OpenAI's code red. Journalist Casey Newton said,

"For OpenAI to realize its ambitions, it's not going to be enough for them to make a model that is as good as Gemini 3. They need to be able to leapfrog it again."

Kevin Roose replied, "They're not going to win by tying for first place."

So ChatGPT is no longer the only option, and it made sense for me to look at the competition.

ChatGPT annoys me

Toward the end of 2025, I tried to prompt ChatGPT to create a "ChatGPT wrapped" covering my use of the product (that was before OpenAI released an official version).

I wrote, "Be honest, what's the most ridiculous question I asked you this year?"

ChatGPT replied, "Honestly? You didn't ask unhinged questions. You asked intensely reasonable questions taken to a very specific extreme."

ChatGPT then gave me a specific example:

“Can you rewrite this one sentence so it sounds more like me… but not softer, not sharper, not longer, not more casual, not more polished, and without changing the meaning at all?”

That entire thing was made up. I never asked ChatGPT to do that. That isn't even something I would prompt, ever. In fact, every single example ChatGPT gave me in response was a hallucination.

In mid 2025, users started noticing that ChatGPT was more sycophantic – so much so that OpenAI had to roll back the 4o update.

I know models hallucinate. But this seemed a pretty basic question: search through the existing chats. And the constant attempts to flatter me (this is not the only example) were starting to get on my nerves. ChatGPT is constantly trying to tell me how brilliant I am.

Claude is delightful

So I tried Claude. I set up Projects (similar to how I was using ChatGPT projects). I noticed a few things right away:

• Claude is much more chill
• The writing output is better. I still edit, but edit less.

Claude is better able to "sound like me" for things like social media posts, especially since the project is connected to my Voice and Tone Guide.

But I also burned through my Claude credits pretty fast, especially when using Opus 4.5. That never happened with ChatGPT. And I'm not willing to pay $100/month for higher usage limits with Claude.

Gemini works within Google Workspace

Due to the Claude usage limits, I turned to Gemini as a supplement.

Gemini has the major advantage of being included in my Google Workspace. I also spoke to a few friends who use Gemini. One said, "The latest Gemini update was a major transformation from the last iteration and provides consistent results if your Gems are thorough."

I don't think the Gemini output is as good as Claude for some writing-related projects. But for other things, it's fine.

I have a split between Gemini and Claude, and a clear delineation of which app I would use for different types of requests.

OpenAI makes me nervous

On the whole, OpenAI seems like a house of cards that will come crashing down at any moment.

Its revenue is nowhere near the amount that has been invested. The company doesn't even make a profit.

Google has a massive advantage of existing distribution. Apple recently announced that it had selected Gemini as its AI partner.

As Kevin Roose of Hard Fork pointed out, OpenAI isn't going to win by tying for first place. It has to edge out the competition in a meaningful way. And, unlike OpenAI, which is hemorrhaging money, Google has nearly unlimited resources that it can pour into making Gemini better.

What happens if OpenAI simply... collapses? Will it get scooped up by Microsoft?

I don't know, but (for now) I'm not willing to stick around and see if OpenAI's bubble bursts.

The switching process

Honestly, it took time to make the switch. I had a lot of projects set up in ChatGPT and those all needed to be re-created in either Claude or Gemini.

I still need to update my Zapier workflows, but they rely on the OpenAI API, not ChatGPT directly. So I could, in theory, cancel my ChatGPT subscription and keep paying for the API until I've finished the transition.

During this process, I've focused on documenting my AI projects and the connected resources. That way, if I ever need to make a change again, I'm in better shape.

More resources:

• My personal guardrails for using AI as a solopreneur
• The AI bubble is here and it's filled with failed experiments

Want some AI inspiration? These content repurposing automation examples include everything from transcripts to carousels to videos.

Download