Claude, ChatGPT, and Copilot all promise to save time, sharpen my work, and make everyday tasks feel effortless. But when each one wants a recurring payment, the real question is: which one actually deserves a place on my credit card bill?
So, I paid for all three for a month and used them the way a normal person would. I used them for writing, research, summaries, emails, planning, coding, and general productivity. After switching between them daily, one service finally felt like a subscription I could defend.
Copilot
Excellent in Office, limited outside Microsoft
Copilot (Microsoft 365 Premium) was the subscription I wanted to like more than I actually did. On paper, it should have been an easy win. It uses GPT models. I already use Microsoft Edge, and OneDrive is part of my workflow.
Microsoft has also been pushing Copilot across Windows and Office as the future of productivity. But after using it for a month, I kept running into the same issue: Copilot is powerful in very specific areas, but limited everywhere else.
For routine tasks like answering questions, providing writing help, summarizing, and brainstorming, Copilot’s answers were solid. The Edge integration was the letdown, though.
I expected Copilot to act like a true browser assistant. Instead, it felt clunky and disconnected. The same applied to OneDrive. I wanted to quickly pull details from documents or presentations, but there is no such option to do so.
Image generation was a highlight. The results were impressive, and I can always refine them in Microsoft Designer. Copilot’s real strength is still Office. In PowerPoint, Excel, and Word, it shines.
I can create an entire presentation using just a prompt, which still feels like magic and isn’t possible in the same way with ChatGPT or Claude. Copilot in Excel and Word also feels robust, especially when I am working inside Microsoft’s ecosystem. However, the UX could have been better, though. The current Copilot icon in the bottom right corner of the Office apps gives me Clippy vibes.
Copilot’s third-party integration is also limited, as it works with Dropbox, Box, and a couple of Google services. I also tried GitHub Copilot in VS Code, and its overall performance is nowhere near that of Claude Code or Codex.
Claude
Brilliant at coding and automation, but an incomplete everyday toolkit
Claude was the closest to making me keep another AI subscription. In several areas, it feels ahead of the pack. Its Fable 5 model is excellent for coding, and until GPT-5.6 becomes widely available, I can see why developers and power users would pick Claude without thinking twice.
When I used it to create a complex website, handle technical tasks, debug ideas, and provide longer explanations, it felt calm, precise, and sharp.
But this test wasn’t about finding the best coding assistant. I wanted to know which AI subscription made the most sense for a regular user like me. That is where Claude became harder to justify.
To be clear, Claude gets a lot right. Its answers are well-structured and rarely feel lazy. The memory feature works well, as I don’t prefer having to repeat my preferences or context again and again.
Projects are useful for keeping work organized, and Cowork makes it feel more like a serious productivity environment than a simple chatbot. The problem is that my workflow is broader than text and code.
I use image generation frequently, whether it is for quick visuals or article ideas. Claude not supporting image generation is a real limitation for me. Yes, I can use Claude Design, but it’s more geared towards creating wireframes, mockups, and slides.
That is why Claude ended up being a near miss. If I were coding heavily every day, I would probably feel differently. But as a casual user with light coding needs, I couldn’t convince myself to keep paying $20 per month for it.
ChatGPT
The most complete AI assistant overall
ChatGPT was the one subscription that finally justified its place on my monthly bill. The entire experience has been almost flawless for my workflow. The answers are spot on; the tone feels natural, and it rarely gives me that robotic response that still shows up in other AI tools.
ChatGPT feels more human, and for a general user, that matters more than raw benchmark numbers. It gives me answers I can actually use without too much editing.
Projects are another big feature. Similar to Claude’s implementation, Projects let me keep related files, instructions, and conversations in one place. For someone who jumps between article drafts, app comparisons, and research notes, that organization makes a real difference.
Codex was also good enough for my needs. I am not a full-time developer, so I don’t need the most advanced coding environment every day. That said, Claude Code is still ahead, and Codex needs to catch up here. I will be watching GTP-5.6 closely here.
It works with all the major third-party services like Canva, Asana, Gmail, and Outlook. My favorite feature is scheduled actions, though. I can ask ChatGPT to run a specific prompt on a specific day and return the results without manually repeating the task.
The image generation has also been top-notch in my case.
My monthly bill needed a winner
It’s easy to understand why ChatGPT wins this comparison for me. Claude is brilliant, especially for coding and deep reasoning. Copilot is excellent inside Office. But ChatGPT is the one that feels useful across most parts of my life. For a casual user with broad productivity needs, it is the only subscription here I can comfortably justify keeping.
