On Thursday, OpenAI launched a chatbot that successfully prices $200 a month, and the AI group had no thought what to make of it.
The corporate’s new ChatGPT Professional plan grants entry to “o1 Professional Mode,” which OpenAI says “makes use of extra compute to get the most effective solutions to your hardest questions. ” Which means. The o1 Professional mode, an improved model of OpenAI’s o1 inference mannequin, ought to be capable to reply science, math, and coding questions extra “robustly” and “comprehensively,” OpenAI stated.
Nearly instantly, folks started requesting to attract unicorns.
I requested ChatGPT o1 Professional Mode to create a unicorn SVG.
(That is the mannequin you’ll be able to entry for $200 monthly) pic.twitter.com/h9HwY3aYwU
— Rammy (@rammydev) December 5, 2024
and design a “crab-based” laptop.
We now have lastly launched o1-pro to the final word use case. pic.twitter.com/nX4JAjx71m
— Ethan Mollick (@emollick) December 6, 2024
And he waxes poetic concerning the that means of life.
I simply signed up for a $200/month subscription to OpenAI.
Please reply to the query. Reposting to this thread. pic.twitter.com/oTQxbPxnoP— Garrett Scott 🕳 (@thegarrettscott) December 5, 2024
Nonetheless, many at X appeared unconvinced that the reply to o1 professional mode was on the $200 stage.
“Has OpenAI shared any concrete examples of prompts that fail on common o1 however succeed on o1-pro?” asked British laptop scientist Simon Willison. “I would wish to see one concrete instance of its advantages.”
That is a good query. In any case, that is the world’s most costly chatbot subscription. The service additionally presents different advantages, equivalent to lifting charge limits and limitless entry to OpenAI’s different fashions. However $2,400 a yr is not a whole lot of change, and the worth proposition stays unclear, particularly for o1 Professional mode.
It did not take lengthy to search out examples of failure. O1 Professional mode struggles with Sudoku, tripped up by optical phantasm jokes which can be apparent to any human being.
Each o1 and o1-pro failed right here, in all probability as a result of visible acuity limitations (additionally on Sudoku puzzles).https://t.co/mAVK7WxBrq pic.twitter.com/O9boSv7ZGt
— Tibor Blaho (@btibor91) December 5, 2024
OpenAI’s inside benchmarks present that o1 professional mode performs solely barely higher than normal o1 on coding and math issues.
OpenAI carried out a “extra rigorous” analysis on the identical benchmark to show the consistency of o1 Professional mode. The mannequin is taken into account to have solved the issue provided that it will get it proper 4 out of 4 instances. Nonetheless, even these assessments didn’t present any dramatic enchancment.

OpenAI CEO Sam Altman as soon as wrote that OpenAI is: path “To an unfathomable and low-cost intelligence.” reveal a number of times Thursday means ChatGPT Professional is not for most individuals.
“Most customers can be very pleased with o1. [ChatGPT] Plastia! ” he stated in X. “Nearly everybody will get essentially the most profit from utilizing the free or plus tier.”
So who’s it for? “Are there actually individuals who pay $200 a month to ask questions like toys?”Write a three-paragraph essay on strawberries without using the letter ‘e’.” or “Please solve this math olympiad problem“?Are they prepared to half with their hard-earned money with little assure that a typical o1 will be unable to reply the identical query satisfactorily?
We requested Ameet Talwalkar, affiliate professor of machine studying. carnegie melon That is his opinion as a enterprise companion at Amplify Companions. “Rising the value 10x looks as if an enormous danger to me,” he informed TechCrunch by way of e mail. “I believe we’ll get a greater sense of the demand for this function in only a few weeks.”
Man van den Block, a pc scientist at UCLA, was extra blunt in his evaluation. “I do not know if this worth is affordable, and I do not know if costly inference fashions will develop into the norm,” he informed TechCrunch.
o1 is “higher than most people at most duties” as a result of, sure, people solely exist in an amnesiac, disembodied multi-turn chat interface. https://t.co/zbLY2BG5pQ
— Aidan McClau (@aidan_mclau) December 6, 2024
To be beneficiant, this can be a advertising failure. Explaining that o1 Professional mode is the most effective at fixing “the hardest issues” would not actually go over nicely with prospects. Neither ambiguous statement About how fashions can “suppose longer” and show “intelligence”. As Willison factors out, with out concrete examples of this purported improved performance, it is arduous to justify paying extra, not to mention 10 instances the value.
This is a extremely attention-grabbing advice immediate for an AI mannequin that prices $2,400 a yr
I hope openai retains these canned pattern prompts as much as ASI pic.twitter.com/JQ5vLKxWWR
— Dean W. Ball (@deanwball) December 6, 2024
So far as I do know, the audience is specialists of their discipline. OpenAI stated it plans to grant free entry to ChatGPT Professional, together with o1 Professional mode, to a small variety of medical researchers at “main establishments.” Errors are crucial in drugs, says Bob McGrew, former chief analysis officer at OpenAI. noticed In X, improved reliability might be the primary unlock for o1 professional mode.
I have been taking part in round with o1 and o1-pro a bit.
They’re excellent and a bit uncommon. Additionally, most often, they don’t seem to be appropriate for most individuals. To extract worth from it, you want a selected arduous downside to unravel. However when you’ve got these issues, this can be a very huge downside.
— Ethan Mollick (@emollick) December 5, 2024
McGrew too lost in thought The o1 professional mode is an instance of what he calls “intelligence overhang.” In different phrases, the elemental limitations of a easy text-based interface don’t present how customers (and maybe mannequin authors) can derive worth from the “further intelligence”. As with OpenAI’s different fashions, the one approach to work together with o1 Professional Mode is thru ChatGPT, which McGrew factors out shouldn’t be good.
Nonetheless, it is also true that $200 raises expectations. Judging by its early status on social media, ChatGPT Professional is something however nice.

