OpenAI has announced a significant shift in its AI development strategy, with CEO Sam Altman revealing plans to release two new models, o3 and o4-mini, in the coming weeks. This move comes as the company delays the much-anticipated GPT-5 model by several months to focus on making substantial improvements. The decision to release o3, previously canceled, and introduce o4-mini, reflects OpenAI's response to technical integration challenges and the need to prepare for unprecedented demand expected for GPT-5. The new models are expected to address specific needs within the AI community. The o3 model, known for its high performance in scientific and mathematical domains, will include a 'pro mode' for handling complex queries. Meanwhile, o4-mini is designed to offer a more accessible option for users. These releases are part of OpenAI's strategy to provide differentiated service levels, with ChatGPT Plus customers gaining access to a higher level of intelligence and ChatGPT Pro subscribers enjoying an even higher level of intelligence.