Introduction
Strawberry has hit the market! No, we’re not talking about the fruit. It’s OpenAI’s latest model, OpenAI o1, codenamed Strawberry. After eagerly awaiting GPT – 5, OpenAI has now released this fact – checking and high – reasoning model, marking a significant leap in AI capabilities, especially in reasoning and problem – solving.
Overview of the OpenAI o1 Model Series
The introduction of the OpenAI o1 model is a game – changer in the world of language models. Unlike its predecessors, it comes with advanced reasoning skills, allowing for more accurate handling of complex queries. There are two variants available for access currently: OpenAI o1 Preview and OpenAI o1 – mini. The Preview model is the more advanced one, enabling deeper analysis and problem – solving, while the o1 – mini is a more budget – friendly option, being around 80% cheaper than the o1 – preview. Both models feature a 128k context window, which means they can process a large amount of information in a single query.
How to Access OpenAI o1/Strawberry?
There are different ways to access the new model, depending on whether you are a ChatGPT user or a developer.
For ChatGPT Users
ChatGPT Plus and Team users can start using the “o1” models right away. There are two models to choose from: o1 – preview and o1 – mini. To select the models, go to the model picker in ChatGPT and manually choose either one. The o1 – preview has a limit of 30 messages per week, while the o1 – mini allows 50 messages per week. OpenAI plans to increase these limits and also enable ChatGPT to automatically select the most suitable model for a given prompt. ChatGPT Enterprise and Edu users will get access to both models starting next week.
For Developers
Developers who are in the API usage tier 5 can start using the “o1” models via the API. Initially, the rate limit is set at 20 requests per minute (RPM), with plans to increase it after further testing. However, the API for “o1” models currently does not support function calling, streaming responses, system messages, and other advanced features. Developers should review the API documentation on OpenAI’s website to understand how to integrate and use these models.
Additional Information
OpenAI is constantly working on enhancing access by increasing message limits and adding features like automatic model selection in ChatGPT. Since the details of access might change, it’s advisable to check the official OpenAI resources for the latest updates.
OpenAI o1 Pricing Model
The o1 – preview is a new reasoning model for complex tasks that require broad general knowledge. It supports up to 128K tokens and has a knowledge cutoff of October 2023. The pricing for o1 – preview is $15.00 per 1 million input tokens and $60.00 per 1 million output tokens. The o1 – mini, on the other hand, is a fast and cost – efficient model tailored for coding, math, and science use cases. It also supports up to 128K tokens with a knowledge cutoff of October 2023, and is priced at $3.00 per 1 million input tokens and $12.00 per 1 million output tokens.
Conclusion
The OpenAI o1 series is a significant step towards more human – like reasoning in AI. Its ability to process information more deeply can lead to major advancements in various fields. As OpenAI continues to improve these models, future versions may further enhance their reasoning capabilities, potentially revolutionizing the way AI systems operate.
