According to CEO Sam Altman, the next version of OpenAI’s artificial intelligence, GPT-5, is set to be released in the coming months. Before then, OpenAI will release GPT-4.5, also known as Orion, which will be the final model without advanced “chain-of-thought” reasoning.
Altman recently released updates on social media, explaining OpenAI’s plans to combine the O-series and GPT-series into a single, more powerful AI system. This connection will allow the technology to think more effectively, adapt its logic to tasks, and provide greater functionality across several applications.
“Our focus is on creating a system that can effectively use all our tools, determining when deeper reasoning is necessary and providing high-quality results across a wide range of tasks,” Altman told me.
One important change in GPT-5 is that OpenAI will no longer offer the O3 reasoning model as a standalone product. Instead, its capabilities will be integrated into GPT-5, increasing accuracy and decreasing disinformation. OpenAI earlier released a smaller version, O3-mini, which aimed to improve reasoning skills in arithmetic, coding, and scientific analysis.
While Altman could not specify a specific release date, he did indicate that GPT-4.5 would be accessible in the coming weeks, with GPT-5 following shortly thereafter. Notably, GPT-5 will be available to free users, albeit with restrictions to avoid misuse. Premium and pro users will have access to a more advanced version with additional features.
The AI sector is very competitive, with alternative models such as DeepSeek shaking up the market. OpenAI’s new release plan indicates a change toward making AI tools more freely available in order to maintain leadership in the quickly changing AI landscape.
