Reasoning models are at the top of the relational AI pack, giving attention to completely thinking through prompts. On March 20, OpenAI revealed the heavy-duty variation of the argument model known as o1-pro.
Starting today, OpenAI API users who are on any of OpenAI’s paid levels may access o1-pro. Higher use tiers may have stricter batch queue and request limits.
The cost reflects higher compute power: o1-pro costs$ 600 for 1 million output tokens and$ 150 for 1 million input tokens. When using the base model of OpenAI o1, compare that to$ 15.00 for 1 million tokens and$ 60 for 1 million output tokens.
Notice OpenAI Recap: o3 Model Wraps 12 Days of Presentations
What features are available with$ 600,000,000 in currencies?
Due to its affordable rate, the OpenAI o1-pro is one of the fanciest AI models. The Silicon Valley darling’s personal GPT-4.5 takes the king from its API exposure. ( GPT-4.5 costs$ 75 per million input tokens and$ 150 per million output tokens. )
What do you receive with that sum of money? A sizable number of researchers, designers, and other professionals using relational AI for research, medicine, or technologies are likely to benefit from o1-pro. OpenAI offers individual models for translation and moderation, in addition to o1-pro’s reasoning models. People ‘ needs for those can probably be met by a smaller, less expensive unit.
For the price of o1-pro, you receive a 200, 000 perspective glass and a 100, 000 maximum productivity currencies.
With text and image type, OpenAI claimed o1-pro you interpret data from device vision. Yet, it only produces word production. In the API, it even includes:
- calling for a performance.
- Organized Outputs, which guarantee that the woman’s response will be based on the JSON Schema used by the developer.
- Integration with the Reactions API, which allows for the creation of AI providers that can perform searches and connect with the website in several ways.
- Inclusion with the Batch API, which enables sequential requests and raises prices and rates for jobs with turnaround times of 24 hours.
Although earlier models had real-world understanding up to soon 2023, OpenAI has certainly made a public announcement about the precise knowledge cutoff date for the o1-pro.
AI companies vie to elude the “reasoning unit” craze.
In September 2024, OpenAI first referred to o1 as” Strawberry.” It competes with Google’s Gemini 2.0, DeepSeek’s R1, Anthropic’s Claude Sonnet 3. 7, Grok 3, and another argument designs. Meta is working on a” theory of mind reasoning,” but Meta claims that it is an hostile venture to evaluate advanced AI designs.
