As OpenAI battles to maintain a competitive edge amid growing pressure in the market – including from Chinese startup DeepSeek – it’s released details of two upcoming models, including the soon-to-be-released advanced-reasoning GPT-5.

Disruption runs deep: DeepSeek’s recent ascendancy has disrupted the AI landscape with its cost-effective and high-performing models, prompting established players to reassess their approaches. In OpenAI’s case, its response has been a promise to streamline its offering and shift towards more advanced reasoning models.

Keep it simple, Sam: The company’s plans were outlined on X in a post from CEO Sam Altman who announced plans to integrate various models into a unified system, aiming to enhance user experience. "We want AI to ‘just work’ for you," he said. "We realize how complicated our model and product offerings have gotten."

Focus on advanced reasoning: First up, GPT-4.5 is slated to be OpenAI's final model without chain-of-thought reasoning capabilities. This move signifies a shift towards more advanced reasoning models, and an ability to handle complex tasks by generating intermediate reasoning steps. GPT-5 – due for release within the next few months – is set to integrate various technologies, including the previously standalone o3 model, into a unified system.

Less is more: Altman acknowledges the proliferation of models available to choose from has added complexity to the user experience. "We hate the model picker as much as you do and want to return to magic unified intelligence," he said in his post.