In a blog post, AI development company, OpenAI announced that it has “begun training its next frontier model.” It is speculated that this model could be GPT-5. The company has not given an official name to this new model yet. However, from various sources, we can understand that it will follow GPT-4 and show improved performance that brings OpenAI nearer to reaching its aim of creating artificial general intelligence (AGI).
“While we are proud to build and release models that are industry-leading on both capabilities and safety, we welcome a robust debate at this important moment,” the blog post stated.
This model is expected to be used in many generative AI tools such as image generators, virtual assistants, search engines, and ChatGPT which is OpenAI’s main chatbot. The process of training and developing AI systems can be very long-lasting for several months or even years.
However, some reports indicate that GPT-5 might be ready sooner than initially expected.
Mistral AI introduces Codestral, its First-Ever Code Model
The Safety and Security Committee
With a strategic plan, OpenAI has set up a Safety and Security Committee that is led by CEO Sam Altman. Policy experts Aleksander Madry (Head of Preparedness), Lilian Weng (Head of Safety Systems), John Schulman (Head of Alignment Science), Matt Knight (Head of Security), and Jakub Pachocki (Chief Scientist) will be a part of the group as well.
This group will dedicate itself to evaluating and improving safety processes in OpenAI for the next 90 days. This move is made ahead of time because the company is expecting possible impacts and dangers related to releasing an AI model that is more advanced.
Mistral AI introduces Codestral, its First-Ever Code Model
What is to be expected from the new model?
Even if OpenAI has not given an official name for GPT-5, it is expected to be released soon. Reports are pointing towards its possible release around mid-year. When asked whether he had tested GPT-5, Satya Nadella first said, “Oh yes, of course.” He did not refute the remark when asked to do so.
GPT-5 will reportedly have better multimodal abilities and start with support for text, image, audio, and video from the beginning. The model may bring new advancements in creating videos from text which could include integrating the latest technologies to enhance response precision and speed. This added function might make GPT-5 a high-tech AI agent that can change many applications in different industries.
Recently, OpenAI released GPT-4o, an improved edition of GPT-4. It can handle many types of input and produce various kinds of output, such as text, audio, images, and video. For audio inputs, it can give answers in just 232 milliseconds. Less costly and quicker in the API, GPT-4o has significant improvements in vision and audio comprehension, rendering it a flexible and cost-effective model.
Top 50+ Sam Altman Quotes on AI for Great Products and Future Technology
