Although it's been over 2 years since the "AI boom" truly began, there's still a wild number of people who exclusively rely on ChatGPT. The funny part is that this isn't because it's objectively the best option. Instead, it's mostly habit, and the fact that most people just never bothered trying alternatives that might actually fit their needs better.

While I've admittedly used ChatGPT more than any other AI chatbot, I finally took the plunge and canceled my ChatGPT (plus Perplexity and Gemini) subscription in favor of Anthropic's Claude.

Claude's coding capabilities are far superior

Kicking things off with the obvious

Although this isn't exactly a surprising point, I've found that Claude consistently produces cleaner, more accurate code than ChatGPT, Gemini, and Perplexity. I've used all four tools extensively on side projects that involve a lot of coding, and I've often found myself in an endless debugging loop with the other three. Once I'd pinpointed exactly what bug was stopping the code from running as expected, I'd ask the AI to fix it. And while it would fix said bug, the new code it generated would typically come with a new set of bugs. With Claude, though, I typically get something workable on my first or second attempt.

In addition to providing code that actually works, Claude also clearly explains the reasoning behind each choice it makes. I've found that Claude understands context a lot better than the alternatives, and when it's unsure about something, the best part is that it actually admits it (or asks you) instead of confidently making things up.

And given we're talking about coding, the biggest reason why Claude has become my go-to is Claude Code. If you haven't used Claude Code before, it's essentially an agentic coding tool that runs directly in your terminal. Unlike chatting with an AI in a browser window and copy-pasting code back and forth, Claude Code actually lives in your development environment. It can see your entire project structure, read your files, understand how different parts of your codebase connect, and make edits directly. It can also run tests, catch errors, and commit changes to GitHub.

Claude is a lot more responsive to feedback

It doesn’t insist it’s right all the time

Something I've noticed consistently with ChatGPT in particular is that it feels like it's always gaslighting me. It rarely backs down when it's wrong, and while it does admit its mistakes, it typically repeats the same flawed logic in a slightly reworded way. Gemini and Perplexity aren't quite as bad, but they still have a tendency to double down rather than genuinely reconsider.

Within days of using Claude, I noticed that it doesn't push back nearly as much. When it's confused or wrong (which let's face it, LLMs often are), I've found that Claude steps back and asks you for your take. This is something I frankly haven't seen any other mainstream AI chatbot do. Given that Claude is a lot more amenable to feedback, it actually implements it, which ultimately makes the final responses a lot more accurate and closer to what you're looking for.

ChatGPT tends to agree with you, even when it shouldn’t

It doesn’t make the same mistakes repeatedly

While it's definitely nice to have someone agree with everything you say, it quickly becomes frustrating when that "agreement" comes at the expense of accuracy or helpfulness. With ChatGPT in particular, you'll constantly come across users complaining about the fact that it feels overly flattering and glazes you even when it shouldn't. In fact, this problem got so bad with a new model update that OpenAI had to roll it back entirely. Gemini and Perplexity are less guilty of this, but they still lean toward telling you what you want to hear rather than what you need to hear.

Although this is something you can sort of deal with by explicitly prompting ChatGPT or adding custom instructions, I think it's a major limitation that you shouldn't have to work around in the first place. This ties back to what I said earlier — when Claude isn't aware of something, or you ask it a question outside its capabilities, it's more likely to say so rather than generating a potentially incorrect response in hopes of pleasing you. That honesty makes a huge difference, especially when accuracy matters, and it's one of the main reasons I've found Claude far more reliable for complex or nuanced tasks.

Claude plays nice with your other apps

The best part

Another reason why Claude is superior to ChatGPT is how easily it connects to other tools I use every day. Claude can connect to hundreds of external tools, and thanks to MCP (Model Context Protocol), it can actually do things within these tools. This includes hundreds of tools like Google Drive, GitHub, Slack, Canva, and more. While some companies have official integrations built directly into Claude, the real magic is that MCP is an open standard. Ultimately, the community keeps building new connectors all the time.

For instance, I stumbled across a NotebookLM MCP server someone developed that lets me connect Claude Desktop directly to my NotebookLM notebooks! This means Claude can see my notebooks, interact with the sources I've uploaded, and generate responses based on that data — all without me ever having to leave the Claude interface.

As someone who extensively relies on NotebookLM and Claude both, the ability to connect the two tools has been transformative. I can create new notebooks, add sources, trigger Audio Overviews, and even query my notebooks directly through Claude. With ChatGPT, Gemini, and Perplexity, you're mostly limited to whatever integrations each company decides to build or approve. With Claude, if an integration doesn't exist yet, chances are someone's already building it (or you can build one yourself).

Artifacts is a feature every AI chatbot should steal

Most underrated Claude feature

With ChatGPT, Gemini, and Perplexity, when you're using them to create a report or help you code something, you always need to copy the output and then paste it somewhere else to actually see or use it. It's a constant back-and-forth, and Claude's Artifacts feature fixes that.

Instead of just dumping everything into the chat, Claude creates a dedicated window right next to your conversation where it displays what it's building in real time. This effectively saves you from needing to jump across multiple tools, and lets you review, tweak, and iterate on the output without ever leaving the interface. It's one of those features that sounds small but completely changes how you work once you start using it.

It might be time to move on from ChatGPT, Gemini, and Perplexity

While ChatGPT, Gemini, and Perplexity certainly have their strengths, it's high time people stopped assuming it's the best option by default. I have zero regrets about switching to Claude, and at this point, I can’t see myself ever going back.