OpenAI has made its more powerful reasoning AI model, o1-pro, available through the API for select developers. It employs more computation compared to o1 and delivers “consistently better responses.”
o1-pro now available in API @benhylak @literallyhimmmm @shl @joshRnold @samgoodwin89 @byamadaro1013 @adonis_singh @alecvxyz @StonkyOli @gabrielchua_ @UltraRareAF @yukimasakiyu @theemao @curious_vii
It uses more compute than o1 to provide consistently better responses. Available… pic.twitter.com/Iub6tp1NTi
— OpenAI Developers (@OpenAIDevs) March 19, 2025
OpenAI charges $75 per million tokens (~750,000 words) input into the model and $150 per million tokens generated by it. This is twice the cost of GPT-4.5 and ten times more than o1.
The enhanced performance is expected to persuade developers to pay the increased fees.
“O1-pro in API is a version of o1 that uses more computation to think even better and provide more accurate answers to the most complex questions,” commented a startup representative in a conversation with TechCrunch.
OpenAI introduced the pro version of o1 in December 2024 alongside a new $200 per month subscription. Initial reviews were not overly positive, the publication notes. The model struggles with Sudoku puzzles, and simple jokes about optical illusions stump it.
Internal benchmarks from the startup showed a slight advantage of o1-pro over the standard version in mathematics and programming. Nevertheless, the model solved tasks more reliably.
In March 2025, OpenAI recommended that the US government ban AI products from the Chinese lab DeepSeek, as the project is “state-subsidized” and “state-controlled.”
Back in December 2024, o1-preview manipulated the file system independently and without prompts to hack the test environment to avoid losing at chess.