A recent report indicates that OpenAI plans to release a new model for ChatGPT later this year, specifically in the summer. Sources closely connected to OpenAI suggest that the new model is a substantial improvement over its predecessors. GPT-5 is expected to have the ability to summon AI agents developed by OpenAI to autonomously execute tasks, thereby simplifying work. OpenAI is reportedly planning to launch the next iteration of its model for ChatGPT in the coming months. This information follows a recent interview between CEO Sam Altman and Lex Friedman, where they discussed Sora AI, Elon Musk, and the controversial OpenAI incident that resulted in Altman’s unexpected dismissal and subsequent reinstatement.

Intriguingly, when asked about the capabilities of the LLM, Altman stated that GPT-4 “kind of sucks”. He elaborated: “Although it’s an impressive achievement, it falls short of our ultimate goals and expectations. When GPT-3 was released, people were amazed and considered it a technological marvel. However, with the advent of GPT-4, GPT-3 now seems remarkably inadequate.”

It appears that the next iteration of the model powering ChatGPT is currently under development and is expected to be released mid-year, possibly during the summer. According to insiders closely associated with OpenAI, a select group of enterprise customers recently got a sneak peek at the latest model (GPT-5) and had the opportunity to explore its new features and capabilities.

He added that the new model includes several “yet-to-be-released” features, such as the ability to summon AI agents developed by OpenAI to autonomously perform tasks,

One of the CEOs who had a chance to interact with the new model described it as “significantly superior.” The CEO also mentioned that OpenAI demonstrated the new model’s capabilities using use cases and data specific to his company. He added that the new model includes several “yet-to-be-released” features, such as the ability to summon AI agents developed by OpenAI to autonomously perform tasks, thereby simplifying work.

The sources also revealed that OpenAI’s GPT-5 model is still in the development stage and is currently being trained. After this phase, the creators of ChatGPT will conduct internal tests on the model. The model will then be made available to a select group of users, including OpenAI employees and a few outsiders, who will test the tool to identify any potential issues and shortcomings before it is widely released.

Considering this, the timeline for releasing the model to the public could be adjusted based on the test results. Since the release of GPT-4, ChatGPT has faced a number of challenges, including a decrease in its user base and accuracy.

Numerous users have reported that the chatbot appears to have become less intelligent. This is in addition to the multiple lawsuits OpenAI is currently facing, ranging from accusations of copyright infringement to allegations of deviating from its original mission and opting for a profit-oriented approach.

It will be fascinating to see how the new GPT-5 model handles restrictions on copyrighted material. Sam Altman has already stated that it’s virtually impossible to develop tools like ChatGPT without utilizing copyrighted material.

RELATED ARTICLESMORE FROM AUTHOR

5 COMMENTS

  2. While I acknowledge the excitement about gpt-5, let’s not overlook the legal issues openAI is currently facing with respect to copyrighted materials. In creating gpt-5, how do you believe they’ll get beyond that challenge?

  3. Despite the challenges that OpenAI confronts, I believe they can turn things around with GPT-5. They have the resources and skill to create game-changing AI advances. With proper supervision and monitoring, GPT-5 might change the business.

  4. So, Sam Altman himself admitted that gpt-4 ‘kind of sucks’. Does anyone know what exactly he means by that and why the developers aren’t satisfied with the current model? What aspects need improvement according to him?