OpenAI unveils most expensive AI model

OpenAI has unveiled the company’s most expensive artificial intelligence (AI) model. The new AI model called “o1-Pro” is reportedly an enhanced version of its o1 "reasoning" AI model.

According to the company, o1-Pro utilises increased computing power to deliver better responses than its predecessor.

Initially, o1-Pro will be available to select developers who have spent a minimum of $5 on OpenAI API services.

However, this premium offering comes with a hefty price tag. The company will charge $150 per million tokens (approximately 750,000 words) fed into the model and $600 per million tokens generated by the model.

This pricing strategy has positioned o1-Pro as a high-end solution, with costs twice that of OpenAI's GPT-4.5 for input and 10 times that of the standard o1 model.

The company is confident that o1-Pro's improved performance will justify the additional expense for developers seeking more reliable and accurate responses.

“o1-pro in the API is a version of o1 that uses more computing to think harder and provide even better answers to the hardest problems,” OpenAI spokesperson stated.

"We're excited to bring this enhanced model to our developer community, offering more reliable responses to their most challenging queries," the spokesperson noted.

Although o1-Pro has been available to ChatGPT Pro subscribers since December 2024, several users have reported that the model struggles with Sudoku puzzles and sometimes tripped up by simple optical illusion jokes.

Moreover, internal benchmarks from  last year revealed that o1-Pro performed only marginally better than the standard o1 model on coding and math problems.