OpenAI has introduced a more advanced version of its o1 “reasoning” AI model, called o1-pro, in its developer API.

According to OpenAI, o1-pro uses more computing than o1 to provide “consistently better responses.”

Moreover, it is only available to select developers — those who’ve spent at least $5 on OpenAI API services — and it’s pricey. Very pricey.

OpenAI is charging $150 for every million tokens (about 750,000 words) input into the o1-pro model and $600 for every million tokens it generates.

This is double the cost of GPT-4.5 for input and ten times more than the regular o1 model.

OpenAI believes that the better performance of o1-pro will persuade developers to pay these high prices.

An OpenAI representative told TechCrunch that o1-pro uses more computing power to provide improved answers to tough questions, and they’re excited to offer it in the API after many developers requested it.

However, initial feedback on o1-pro, which has been available to ChatGPT Pro users since December, hasn’t been very positive.

Users reported that they struggled with Sudoku puzzles and had trouble with simple optical illusion jokes.

Additionally, some internal tests from late last year showed that o1-pro only performed slightly better than the standard o1 on coding and math problems, although it did provide more reliable answers.

Read more

ChatGPT search opens to all users in challenge to Google

ChatGPT canvas introduced for enhanced collaboration in writing and coding

OpenAI chief Altman says DeepSeek’s R1 model ‘impressive’