OpenAI has introduced an upgraded version of its reasoning AI model, known as o1-pro, now available through its developer API. This new model promises to deliver improved performance by leveraging greater computational resources compared to its predecessor, o1.
Currently, o1-pro is accessible only to a select group of developers—specifically those who have spent a minimum of $5 on OpenAI API services. However, this advanced model comes at a steep price, making it a significant investment for those interested in utilizing its capabilities. OpenAI has set the costs at $150 per million tokens (approximately 750,000 words) for input into the model, and $600 per million tokens for the output generated. These rates are considerably higher than those for OpenAI’s GPT-4.5, with input costs being twice as much and output costs ten times higher than the regular o1 model.
OpenAI is banking on the enhanced functionality of o1-pro to justify its premium pricing. A spokesperson highlighted that this version is designed to “think harder” and deliver more reliable answers to complex problems, a response to numerous requests from the developer community for improved capabilities.
Despite its promising features, initial feedback on o1-pro has been mixed. Although it has been available for ChatGPT Pro subscribers since December, users reported challenges with tasks like solving Sudoku puzzles and understanding simple optical illusion jokes. Moreover, internal benchmarks from OpenAI late last year indicated that while o1-pro showed slight improvements over standard o1 in coding and math problems, the differences were not as significant as anticipated. Nevertheless, it was noted that o1-pro provided more consistent answers in these areas.
As OpenAI continues to refine its offerings, the introduction of o1-pro marks a notable step in the evolution of AI reasoning models, inviting developers to weigh the benefits of enhanced performance against the costs involved.
