OpenAI's super-powerful inference model o1 and ChatGPT Pro: 10 cases to take you through!

In the history of Artificial Intelligence, September 12, 2024 will perhaps be remembered as a historic turning point. It was on this day that OpenAI announced its newest series of AI models - o1. According to officials, o1 is the smartest model ever and thinks closer to humans, marking a new path to the age of reasoning that humanity is on.

On December 6, 2024, OpenAI released the o1 model, and the model released this time is smarter than the o1-preview model, especially in terms of programming capabilities, and the ability to adjust the response speed according to the difficulty of the answer, and also added the ability to recognize pictures. The company also launched the paid version of ChatGPT Pro, which is $200/month, and this version has no time or frequency limit.

In just 3 days, OpenAI's o1 and Pro paid versions have blown up in the AI community.

Someone thanked, "$200/month is really expensive, but after trying Pro, it's only going to be like 'it's worth the money!"

Today, we're going to take a look at the following 10 cases to see if the "Most Powerful Pro on the Planet" lives up to its name!

1. Read the flowchart and write the code directly

Upload a flowchart directly and o1 Pro will give you the corresponding high-quality code.

While there is still room for improvement in the code, it is already far superior to previous versions.

2. Superlative Scrabble creativity

Someone had the o1 Pro generate an alphabet grid where each 7-letter word fit perfectly into the nicknames of the members of the Fantastic Four.

This ability to play with words is nothing short of entertaining and cerebral, and it's a straight-up blast.

3. Write the Coinbase clone code in one line of prompts

Someone got o1 Pro to jack the Coinbase clone code with just a one-sentence hint word.

Screenshots + a round of comparisons of 4 popular models show that the o1 Pro hangs all over the place.

o1 Pro 

Other models 1

Other models 2

Other models 3

4. Masters of writing style

More than just math and code, the o1 Pro is pretty amazing at writing.

It takes on everything from serious academic debates to rhyming poems, jokes, and even movie scripts with precision.

5. From a cartoon to complete software

Someone had the bright idea of using one of XKCD's comics as a prompt to make o1 Pro a fully functional software.

The result? 15 minutes of work, with all the features + GUI in place.

6. Math Ceiling: Solving IMO Contest Problems

The third problem of the 2006 Math Olympiad was said to have been tackled by only 28 competitors worldwide that year, and it took 4.5 hours.

And the o1 Pro only took 6 minutes and 48 seconds and explained each step clearly, which is fierce!

7. Solve Sudoku in seconds

Sudoku fans gather! o1 Pro not only solves complex sudoku, but also generates puzzles and the corresponding steps to solve them.

Maybe all the sudoku puzzles you play in the future will be from the AI.

8. Complex calculations? Piece of cake.

"x - 14 to the 10th power, x being the number of vowels in the answer."

With such a tricky problem, the o1 Pro not only does the math, but also explains the steps in detail.

9. Breaking the "impossible" Connections puzzle

Two weeks ago, someone posted that no LLM could solve the NYT Connections puzzle.

But o1 Proo1 Pro not only can solve, but also accurate every time, directly face the conclusion of the study, the speed of AI progress is really so fast that people are afraid.

10. An annotated chronology of scientific and technological developments from 1990 to the present.

o1 Pro was asked to prepare an annotated chronology of science and technology from 1990 to the present, which turned out to be informative and logical.