Sign up to save tools and stay up to date with the latest in AI
bg
bg
1

OpenAI's o1-pro is the company's most expensive AI model yet | TechCrunch

Mar 19, 2025 - techcrunch.com
OpenAI has introduced o1-pro, an enhanced version of its o1 "reasoning" AI model, available through its developer API. This model, which requires more computing power, is designed to deliver consistently better responses than its predecessor. However, access is limited to developers who have spent at least $5 on OpenAI API services, and the pricing is steep—$150 per million tokens for input and $600 per million tokens generated, significantly higher than other models like GPT-4.5. OpenAI believes the improved performance of o1-pro will justify the cost for developers seeking more reliable answers to complex problems.

Despite its promise, early feedback on o1-pro has been mixed. While it has been available to ChatGPT Pro subscribers since December, users have noted its struggles with tasks like Sudoku puzzles and optical illusion jokes. Internal benchmarks from OpenAI indicate that o1-pro performs only marginally better than the standard o1 on coding and math problems, though it does so more reliably. The model's reception highlights the balance between enhanced capabilities and the high cost associated with its use.

Key takeaways:

  • OpenAI has launched o1-pro, a more powerful version of its o1 AI model, available to select developers through its API.
  • O1-pro is significantly more expensive than previous models, costing $150 per million tokens for input and $600 per million tokens generated.
  • Early impressions of o1-pro have been mixed, with some users noting struggles with certain tasks like Sudoku puzzles and optical illusion jokes.
  • Despite mixed reviews, OpenAI claims o1-pro provides more reliable responses and is betting on its improved performance to justify the higher cost.
View Full Article

Comments (0)

Be the first to comment!