Why I Decided to Run This Test
I write for a living. Emails, blog posts, even the occasional love note to my wife (she says I need help). So when people ask me which AI assistant is actually worth paying for, I don’t have a quick answer. Not anymore. The landscape has changed since 2024.
Back then, ChatGPT was the obvious choice. Then Claude showed up with its “constitutional AI” thing, and Gemini started flexing Google’s search muscle. By June 2026, all three are mature products. But they’re not the same. Not even close.
So I ran a little experiment. I gave each one the same five writing tasks—a marketing email, a technical explanation, a creative story, a persuasive essay, and a casual blog post. I judged them on tone, accuracy, creativity, and how much I had to edit afterward. Here’s what happened.
Round 1: The Marketing Email
I asked for a short email promoting a new productivity app called “FocusFlow.” Target audience: busy professionals. Tone: friendly but professional.
ChatGPT 5.0 came out swinging. It wrote a clean, direct email with a subject line that actually made me want to open it: “Stop multitasking. Start doing.” The body was concise, highlighted three key features, and ended with a clear call to action. I’d send that to a client without changing a word.
Claude 4 took a different approach. It wrote a longer email, almost like a mini-essay about the philosophy of focus. It was thoughtful and well-researched, but honestly? Too wordy for a marketing email. I’d have to cut half of it.
Gemini 3 struggled here. It wrote a perfectly fine email, but it felt generic. Like it had read a hundred marketing emails and averaged them out. No personality. No punch.
Winner: ChatGPT. It understood brevity and persuasion.
Round 2: The Technical Explanation
I asked for an explanation of how blockchain works, aimed at a general audience. No jargon.
Claude was the clear winner. It started with an analogy about a shared notebook, then gradually built up the concept without ever getting confusing. I actually learned something, and I’m not a tech person. The explanation was patient and clear.
ChatGPT was accurate but drier. It felt like reading a textbook. Gemini tried to be concise but skipped over important context. If I were a beginner, I’d be lost.
Winner: Claude. For teaching, it’s unbeatable.
Round 3: The Creative Story
I asked for a 500-word story about a robot who learns to feel emotions. This was the most fun.
ChatGPT wrote a story that was clever and well-structured, but emotionally cold. The robot’s journey felt mechanical (pun intended). Claude went the opposite direction—almost too emotional, with a melodramatic ending that made me roll my eyes.
Gemini surprised me. It wrote a story that was genuinely moving, with a quiet, understated tone. The robot didn’t cry or declare love. It just made a small choice that showed it cared. That hit harder.