OpenAI to launch GPT-5 model in August: What to expect
What's the story
OpenAI is gearing up to launch its next-generation language model, GPT-5, as early as next month. The news comes after Microsoft engineers were reported to have been preparing server capacity for the new model. OpenAI CEO Sam Altman also recently teased the upcoming release on X, saying "we are releasing GPT-5 soon."
Model showcase
Altman recently demonstrated the capabilities of GPT-5
In a recent podcast with Theo Von, Altman demonstrated the impressive capabilities of GPT-5. He asked the model a question he didn't understand and was amazed by its instant, accurate response. "I felt useless relative to the AI," he said, describing it as a "weird feeling." This demonstration has further fueled speculation about an imminent launch of this next-generation language model.
Model variants
GPT-5 will be available in 3 versions
OpenAI is expected to launch GPT-5 in the coming days, along with mini and nano versions also available via its API. Altman had previously described GPT-5 as "a system that integrates a lot of our technology." The model will include o3 reasoning capabilities, which were initially meant to be shipped separately. This move is part of OpenAI's efforts to simplify and combine its large language models into a more capable system.
AGI ambition
Development of GPT-5 toward achieving artificial general intelligence
The development of GPT-5 is a step toward OpenAI's goal of achieving artificial general intelligence (AGI). This milestone is crucial for the company as it would force Microsoft to give up its share of OpenAI profits and future AI models. However, Altman has said that GPT-5 won't have a "gold level of capability for many months" after launch.
Open LLM
OpenAI is also working on an open-weight model
While the launch of GPT-5 is expected in early August, OpenAI's release dates are often subject to change due to development challenges or server capacity issues. The company is also preparing an open language model, which reports say is "similar to o3 mini," with reasoning capabilities. This would be the first time OpenAI has released an open-weight model since GPT-2 in 2019.