OpenAI has recently launched its latest AI model, the o1-pro, designed to deliver enhanced reasoning capabilities. This model is accessible through the developer API and represents a more powerful iteration of OpenAI's o1 model. It leverages significantly more computational resources to ensure consistently superior responses to complex problems.

The announcement was made via a post on X (formerly Twitter), stating that o1-pro is now available in API to select developers on tiers 1–5. The post highlighted key features such as support for vision, function calling, structured outputs, and compatibility with the Responses and Batch APIs. The increased computational power comes at a higher cost: $150 per 1 million input tokens and $600 per 1 million output tokens.

Priced at $150 per million tokens for input and $600 per million tokens for output, the o1-pro model is twice as expensive as OpenAI's GPT-4.5 and ten times the cost of the standard o1 model. OpenAI believes that the model's enhanced performance justifies the higher price, particularly for developers working on challenging tasks.

The o1-pro model includes a 200,000-token context window, supports image inputs, and offers structured outputs, making it suitable for applications that require precise and reliable responses. Early benchmarks indicate that while o1-pro shows slight improvements over its predecessor in coding and math problems, the advancements are incremental.

Currently, o1-pro is available only to select developers who have spent at least $5 on OpenAI's API services. It is primarily targeted at AI agents—applications designed to perform tasks autonomously—through OpenAI's new Responses API. The Chat Completions API does not currently support access to o1-pro.