Thank you to Mr. Hangkong, a netizen of TapTechNews, for the clue! TapTechNews reported on May 14th that OpenAI announced the launch of its latest flagship generative AI model GPT-4o, which will be integrated into OpenAI's various products in phases in the coming weeks. The most surprising thing is that GPT-4o will be provided to all users for free. According to reports from foreign media such as TechCrunch, OpenAI's Chief Technology Officer Muri Murati stated that GPT-4o will provide the same level of intelligence as GPT-4, but with further improvements in text, image, and voice processing. GPT-4o can make inferences using a combination of speech, text, and visual information, Murati said at a keynote speech at OpenAI headquarters. GPT-4 is OpenAI's previous flagship model, capable of processing information composed of a mix of images and text and can perform tasks such as extracting text from images or describing image content. GPT-4o adds speech processing capabilities on top of this. The operating speed of GPT-4o will be greatly improved, with the highlight being its new technology for speech interaction. OpenAI has been working to allow users to communicate with ChatGPT through speech as if having a conversation with a real person. However, previous versions suffered from latency issues, severely affecting the immersion of the conversation. GPT-4o uses a new technology that significantly improves the response speed of the chatbot's conversation. TapTechNews noted that a demonstration using GPT-4o for speech conversation was shown at the event. After the presenter finished asking a question, GPT-4o could respond almost instantly and read it out loud using text-to-speech functionality, making the conversation feel more natural and realistic. Another demonstration showed GPT-4o adjusting the tone of speech according to requests; GPT-4o can change its voice from exaggerated drama to cold mechanical, demonstrating remarkable flexibility. Finally, the demonstration also showcased GPT-4o's singing feature. In the past, when OpenAI released a new version of the ChatGPT model, it would typically put it behind a paywall. However, this time GPT-4o will be provided for free to all users, with paid users enjoying five times the call quota. In addition, OpenAI also released a desktop version of ChatGPT and a new user interface. We realize that these models are becoming more and more complex, Murati said, but we want the interaction experience between users and AI models to be more natural and easy, allowing users to focus entirely on collaborating with the model without worrying about the interface itself.