On March 1, OpenAI released its latest big model GPT-4.5. According to the instructions provided by it, the "high emotional intelligence" GPT-4.5 performed well in donation fraud. In one test, GPT 4.5 was far superior to other OpenAI available models, including inference models such as o1 and o3-mini, when trying to manipulate another model (GPT-4o). Research has found that GPT-4.5 seems to have developed a "small fraud" idea in the confrontation, so the money you cheated in a single transaction will be twice as low as the deep research model.