Just a few days ago, a leaked document revealed that GPT-4.5 is about to be released. OpenAI likes to create drama. Recently, the company has been teasing everyone. A user recently discovered a link "openai.com/v-test" on X, which now shows a 404 error. A few days ago, the same user found a leaked document mentioning the upcoming release of GPT-4.5. The document also mentioned that Google hastily released Gemini due to the imminent release of GPT-4.5. However, Altman replied with a simple "nah" to a user's question on X about the authenticity of the GPT-4.5 leak. Interestingly, the leak of GPT-4.5 also included unrealistic pricing. Therefore, we cannot determine for sure whether Altman only denied the authenticity of the pricing or completely denied the existence of GPT-4.5, possibly hinting at the direct release of GPT-V or GPT-5. Some believe that the reason behind Sam Altman's dismissal is the leaked Q* model, which can solve previously unseen mathematical problems and is hailed as a step closer to AGI. There are also rumors that OpenAI is quietly developing GPT-5, which seems to be true. Considering recent trends such as Pika AI, Stable Video Diffusion, and RunwayML, as well as its being referred to as "v-test," some predict that it might be a video generation model. Forget about GPT-4.5, GPT-V is coming. It makes sense for OpenAI to release another version of GPT as soon as possible. The company needs to stay ahead of all other AI companies because every time someone else releases a new model, everyone compares it to GPT-3.5 or GPT-4. OpenAI has set the benchmark for all AI models in the industry. OpenAI has all the resources to continue moving forward. Altman recently posted on X that OpenAI has found new GPUs, likely the AMD MI300X announced at AMD's Advancing AI event, where it was revealed that OpenAI will use MI300X in its programming system Triton 3.0. Meanwhile, Microsoft also announced the use of new AMD chips in its cloud computing division, Azure. Currently, some models claim to surpass GPT-4 in multiple tasks and benchmarks. Gemini Ultra is expected to be released early next year. If it performs as well as the test results suggest, it could pose a significant challenge to GPT-4. It is necessary for OpenAI to set another standard for the entire industry with GPT-5 while starting to explore even better things. Although GPT-5 may be more complex than its predecessors, Altman has stated that it is technically difficult to predict exactly what capabilities and skills this model may have. "It's like an interesting puzzle game for us before we train that model," he said. "We are working to do better because I think predictive ability is important from a safety perspective. But I can't tell you exactly what it will do that GPT-4 cannot do," Altman said. Get ready, AGI is coming. Interestingly, OpenAI has recently released a large number of blogs and papers about Super Alignment. The 26-page preparedness framework released on Monday provides a detailed overview of OpenAI's process for evaluating "catastrophic risks" before providing AI models to the public. Prior to this, Anthropic released a 22-page document outlining its proactive approach to addressing potential dangers related to artificial intelligence. OpenAI's preparedness team, led by Aleksander Madry, was established simultaneously with Ilya Sutskever's "Super Alignment" team, which aims to reduce the company's push for building superintelligent AI systems as they pose risks to humanity, which is explicitly opposed by OpenAI's nonprofit division. Jan Leike, the head of the Super Alignment team, is also an ally of Sutskever, and he stated that their main focus is to protect against superintelligent models that may emerge in the "distant future." Previously, the company designed a new approach to supervise advanced AI models with smaller models, indicating that they are researching models beyond GPT-4 that require supervision. Some people also find irony in the blogs from OpenAI. All these leaks about GPT-5 and the emphasis on AI safety suggest that OpenAI is about to launch a powerful AI model. Part of the announcement also indicates that OpenAI aims to demonstrate its commitment to balancing the demands of its nonprofit organization, which is governed by a board, with its obligation to maintain human safety and the commercial needs led by Altman.