OpenAI to retire GPT-4, launch new AI models next week
What's the story
OpenAI is gearing up to launch a new artificial intelligence model, GPT-4.1, possibly as early as next week.
The Verge reports that the new model is expected to be an improved version of the current GPT-4o.
In addition to the main release, two other variants called GPT-4.1 Mini and Nano are also expected to be launched simultaneously.
Model transition
Future plans and model discontinuation
OpenAI CEO Sam Altman recently had announced plans to release a comprehensive version of its o3 reasoning model, along with the o4 mini.
The Verge suggests that these models could also debut next week.
Ahead of this new launch, OpenAI has confirmed in a recent changelog that it will be discontinuing the GPT-4 model from April 30.
Reasoning
Rationale behind discontinuing GPT-4
OpenAI detailed its decision to retire GPT-4, saying, "GPT-4o is our newer, natively multimodal model. In head-to-head evaluations it consistently surpasses GPT-4 in writing, coding, STEM and more."
The company noted that recent improvements have further enhanced the instruction-following capabilities of the new model.
However, it also assured users that "GPT-4 will still be available in the API."
Enhancements
New features and updates in ChatGPT
In a recent update, OpenAI integrated native image generation capabilities into GPT-4o. The enhancement enables the model to generate more precise and nuanced images.
OpenAI also introduced a new memory feature in ChatGPT that allows the chatbot to remember details about users' previous chats. This way, it can personalize its responses accordingly.
However, users can disable this memory feature if they want.