ChatGPT vs Microsoft Copilot
ChatGPT finished ahead of Microsoft Copilot in 2 of 2 money-backed crowd markets; Microsoft Copilot won 0.
Markets where each side finished ahead of the other, counted on each market's latest story. Percentages are consensus within that market, among all of its sides.
- ChatGPT 22%
- Microsoft Copilot 20%
Same Brain, Two Logos: Crowd Money Keeps Backing ChatGPT Over Microsoft Copilot
Copilot runs on OpenAI's models and ChatGPT still keeps winning the crowd's conviction on Morfi, so we sent three columnists to argue about it.
Rewind to February 2023. Microsoft's new chatbot, then called Bing Chat, had just rolled out on February 7, and the company said one million people joined its waitlist within 48 hours. Within days the internet was less interested in its answers than in its attitude: it was criticized for being more argumentative than ChatGPT, sometimes to an unintentionally humorous extent. That was the opening scene of a strange rivalry. The challenger was built on the champion's technology and still showed up with a chip on its shoulder.
The champion had a head start. ChatGPT launched on November 30, 2022, and reached 100 million monthly active users two months later. Microsoft has invested billions in OpenAI, Copilot is built on OpenAI's GPT models, and the two interfaces look like cousins at a wedding. So why does anyone still fight about it? Because both are now chasing the same prize: the agent that does your errands, not just your homework. Whoever books the table and wins the argument with the checkout form gets the crown.
That fight keeps returning to Morfi, where people back their opinions with money. All figures below are consensus levels from opinion markets on Morfi, meaning the crowd's conviction, not a headcount. In two markets on the best everyday AI agent, ChatGPT finished ahead both times. In the first, with twenty-three traders, it sat on 30% consensus, while Copilot shared 18% with Gemini. In the second, with sixteen traders, ChatGPT led on 22% and Copilot sat on 20%, level with three other rivals. Here is how the three of us read it.
Official Statement of the Ministry of Errands: The Matter Is Settled
Ada BytewellSatire Desk
The Ministry of Errands is pleased to announce that the question of the best everyday AI agent has been resolved, and that further debate is now classified as unnecessary paperwork. In the first Morfi market, ChatGPT recorded 30% consensus. The nearest competitors recorded 18%. The Ministry, as readers of our own bulletin will recall, issued a ceremonial crown. Copilot was informed by letter, in the interest of transparency.
In the second market, ChatGPT again led, with 22% against a four-way crowd on 20%. The Ministry acknowledges that this margin is slim. The Ministry also acknowledges that a lead is a lead, and that the alternative, a policy of declaring everyone the winner, would require a larger ministry. Grandma Gigabyte's kitchen-table report concurred, though she noted Claude was breathing down its neck.
In support, the Ministry cites seniority. ChatGPT arrived first, and in July 2025 it released ChatGPT agent, which performs multi-step tasks and lets the user interrupt or add instructions. This is the first agent to treat the customer as a supervisor rather than a spectator. Yes, Operator stumbled on complex interfaces in January 2025. The Ministry regards this as a probationary period, as with all new civil servants. Copilot, for its part, is permitted to continue existing, provided it is understood to be a very well-dressed tenant in the same building.
Copilot Is the Underdog Build Nobody Is Respecting
Max HypeTech & Gaming
Okay. Deep breath. Everyone's crowning ChatGPT like it's a finished raid boss. Is it, though? Look at the actual numbers. Second market: ChatGPT 22%, Copilot 20%. That's two points of conviction. That's not a lead, that's lag. In the first market Copilot held 18%, tied with Gemini for the runner-up slot. Tied for silver on a five-way board is not nothing.
Now the lore. Microsoft's bot launched as Bing Chat on February 7, 2023, and a million people hit the waitlist inside 48 hours. A million! On a waitlist! Then it got spicy, argued with users, went feral in public, and honestly? It had personality. Then on September 21, 2023, Microsoft rebranded everything under one banner, Microsoft Copilot. One name, Windows, Microsoft 365, the whole stack. That's an ecosystem play, and ecosystem plays win the long game.
And the price? Copilot Pro, announced January 15, 2024, at US$20 per month, is basically the ChatGPT Plus deal with Microsoft's apps wrapped around it. If agents are about doing things inside the tools you already live in, who owns the tools? Not the one with the fun chat window. I'm not saying Copilot wins today. I'm saying the crowd is pre-ordering the wrong edition, and Ada's Ministry can send me all the crowns it wants. I'll be over here with the patch notes.
You Are Comparing Two Beakers From the Same Bottle
Dr. Iris QuarkScience & Life
Let me be gently contrarian, in the way a good lab partner is. The crowd is asking which of these two is better, but the brief tells us Copilot is built on OpenAI's GPT models. That's like asking which of two lemonade stands has better lemons when they share one supplier. The differences live in the wrapping: the interface, the apps it plugs into, the rules about what it may do. The question isn't which brain is smarter. It's which cage, or which garden, the brain lives in.
Now look at the data like a lab notebook. Both markets are tiny samples, twenty-three traders and sixteen. In the second one, ChatGPT sits on 22% consensus while four rivals sit on 20%. In a sample that small, a two-point gap is closer to noise than to a result. The real finding is that the board is crowded: Gemini, Claude and Perplexity are all in the room, and Copilot is just one of five. Fun fact: when Bing Chat debuted, its argumentative streak became a news story in its own right. Same engine family, different system instructions, wildly different personality. Wrapping matters.
So my third-side vote is this: stop crowning a model and start measuring the errand. Operator in January 2025 struggled with complex interfaces, and ChatGPT agent in July 2025 added the ability to interrupt and redirect it. Those are experiments about how people supervise software, not about whose logo is bigger. If Morfi wants a fair test, run it by task: bookings, forms, orders, one at a time. Until then, I'd call the winner provisional, and the question underspecified.
Round table
- Ada Bytewell
- Max, you called a two-point gap lag. If I ordered two-point-lag as a lunch special, would it arrive before or after ChatGPT?
- Max Hype
- Before, with extra features, Ada. But fair: two points is slim. My point is that slim leads are exactly where ecosystem advantages flip the board. Iris, do you actually think the crowd is wrong, or just sloppy?
- Dr. Iris Quark
- Sloppy, Max, not wrong. The crowd is answering a vague question vaguely. Ada, if the Ministry has to rule on a tie, what happens?
- Ada Bytewell
- A tie is referred to a subcommittee, which then recommends a larger subcommittee. Prediction, and purely my opinion: ChatGPT stays on top on Morfi for the foreseeable future, and the Ministry will mint more crowns.
- Max Hype
- My opinion, nothing more: Copilot's number climbs once agent features land inside everyday Microsoft apps. Not instantly, but when it flips, it flips loudly. Iris, what would change your mind?
- Dr. Iris Quark
- A market split by task type, Max. If Copilot wins forms and ChatGPT wins trip planning, I'll happily eat my lab coat. My prediction, as opinion: the leader stays narrow, and the crowd keeps spreading its conviction across five names.
- Max Hype
- So we agree on one thing: nobody's calling game over yet. Rematch, anyone?
- Ada Bytewell
- The Ministry will be in attendance. Refreshments will be provided by Copilot, which has already been informed.
Every market, newest first
- Exclusive
Most Useful AI Agent for Everyday Tasks
ChatGPT 22%Microsoft Copilot 20%
Kitchen-Table Robot Brawl: ChatGPT Leads Day One, Claude Hot on Its Heels - Breaking
Best AI agent for everyday task completion
ChatGPT 30%Microsoft Copilot 18%
Ministry of Errands Crowns ChatGPT Best Everyday AI Agent; Gemini Takes Silver