OpenAI is preparing to launch GPT-5 in early August, according to sources cited by The Verge. The release marks a major step in unifying the company’s language models with its advanced reasoning systems.

OpenAI CEO Sam Altman recently confirmed the upcoming release on X while celebrating a gold medal at the 2025 IMO competition, achieved with a general-purpose reasoning model.

GPT-5 will integrate the o3 reasoning model into a single system, streamlining usage across ChatGPT and developer tools. Altman previously described it as “a system that integrates a lot of our technology.”

Before GPT-5’s debut, OpenAI also plans to release an open-weight model for the first time since GPT-2. This model, similar to o3 mini, will include reasoning capabilities and be made available via Azure, Hugging Face, and other cloud providers.

The announcement comes as OpenAI has rolled out several new features for ChatGPT over the past month. In June, the company introduced agentic capabilities that let users delegate tasks like scheduling, slideshow creation, and code execution.

These agents combine earlier tools like Operator and Deep Research into a single interface, further expanding ChatGPT’s functionality ahead of GPT-5’s release.