Sam Altman, the CEO of OpenAI, has announced the roadmap for the company’s new generation of artificial intelligence models. The goal is to simplify complexity by eliminating the need for users to choose which model to use, instead opting to integrate AI models into a single suite for user convenience.
Currently, OpenAI’s main models available for use are the GPT family, focusing on immediate responses, with the latest being GPT-4, and the O family, emphasizing step-by-step thinking, recently launching O3 and O3-mini. As a result of consolidating these models, OpenAI has discontinued the distribution of the O3 model for users, incorporating its step-by-step process into the GPT model as the main representation.
Moving forward, OpenAI’s plan involves releasing the GPT-4.5 model, internally coded as Orion, followed by the GPT-5, combining the functionalities of the O models. Users will be able to decide whether a question should be approached step-by-step or not.
Altman also mentioned that on the launch day of GPT-5, free ChatGPT users will have immediate access to the latest model. Paying customers, including Plus and Pro tiers, will have access to advanced capabilities beyond the free users.
In conclusion, Altman responded to questions about the timeline, stating that GPT-4.5 will be released in a few weeks, while GPT-5 will debut in a few months. There are also considerations to expand usage limits for heavy users, allowing payment based on processing quantities, especially for the limited ChatGPT Deep Research tool with a cap of 100 queries per day.
TLDR: OpenAI’s new AI model roadmap aims to streamline complexity by integrating different models into a single suite for user convenience, with upcoming releases including GPT-4.5 and GPT-5, offering advanced capabilities for paying customers and expanding usage limits for heavy users.
Leave a Comment