OpenAI Unveils Costly o1-pro AI Model

# openai # o1-pro

OpenAI has rolled out a beefed-up version of its o1 "reasoning" AI model, dubbed o1-pro, through its developer API. They're touting that o1-pro crunches more data than its predecessor, promising "consistently better responses." But here's the catch: it's only accessible to developers who've shelled out at least $5 on OpenAI's API services, and it comes with a hefty price tag. We're talking $150 per million tokens (that's about 750,000 words) you feed into the model, and a whopping $600 per million tokens it spits out. That's double the cost of OpenAI's GPT-4.5 for input and a staggering ten times the price of the regular o1 model. OpenAI is banking on o1-pro's enhanced performance to justify these sky-high prices. "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," an OpenAI spokesperson dished to TechCrunch. "After getting many requests from our developer community, we're excited to bring it to the API to offer even more reliable responses." But don't get too excited just yet. Early feedback on o1-pro, which has been available to ChatGPT Pro subscribers since December, hasn't been all sunshine and rainbows. Users reported that the model stumbled over Sudoku puzzles and got bamboozled by simple optical illusion jokes. What's more, some internal benchmarks from OpenAI late last year revealed that o1-pro only edged out the standard o1 slightly on coding and math problems. Sure, it answered those questions more reliably, but the improvement was pretty modest.

Related article

Comments (15)