VERDICT: Claude Named Most Helpful AI as ChatGPT Is Politely Thanked for Its Time

The Ministry of Helpfulness is pleased to announce that the question of which chatbot is most helpful has been settled. On Morfi's opinion market "ChatGPT vs Gemini vs Claude: Most Helpful AI", Claude finishes in first place with 44% consensus. ChatGPT takes second at 28%. Officials stress that both results were reached by humans, without any help from the contestants, which the Ministry considers a first for the sector.

The candidates need little introduction, so the Ministry will introduce them anyway. ChatGPT was built by OpenAI and first released on November 30, 2022, a date now kept as a public holiday by the essay-writing community. Google Gemini, formerly known as Bard, is ranked by Statcounter as the second-largest chatbot in the world, which means it is big but has not quite won the popularity vote. Anthropic sells Claude-based tools including Claude Code, a terminal coding agent that takes coding tasks in plain English, so developers can now argue with their terminal in full sentences.

The Ministry also looked at the scientific evidence. A 2025 red-teaming study of a couple of hundred patient-posed medical questions, published in a Nature journal, found that Claude gave the smallest share of problematic answers, followed by Gemini and then GPT-4o. The researchers point out that they tested older versions of each model. In AI, older can mean last season. A separate 2024 study of spatial tasks found that all of the chatbots tested did well on GIS theory but struggled with maps, writing code and spatial reasoning. The Ministry therefore advises citizens not to ask any of them for directions to the Ministry.

The verdict comes from 31 traders who put $118,947 behind their opinions in less than a day of trading. Also on the ballot were 1 other side. The crowd's conviction does not quite match the science, but it does lean the same way: the safest performer in the medical study is also the one people backed as most helpful. The Ministry notes that helpfulness is, by the market's own definition, a matter of personal taste, and has decided to treat it as binding law anyway.

In the Ministry's opinion, the only fair response is a Helpfulness Allocation Scheme. Under it, the winning assistant will be assigned all citizens' writing and coding work, the runner-up will take explanations, and Gemini will be put in charge of telling people what the weather was. Any chatbot that fails to rank first will have to start every reply with a short apology, a requirement several of them reportedly meet already. Readers who want the arms-race side of this story should see Max Hype's dead-heat dispatch on AI safety versus speed, which this Ministry has filed under Matters Still Pending.

Next, the Ministry will be watching the model lineups, which change faster than its letterhead. Any of today's rankings could be overturned by a new release before this notice has finished printing. The market is still open on Morfi, and citizens who disagree with the verdict can put money behind their objection there. That is still the only complaints procedure the Ministry recognises.