Recently, OpenAI announced on the X platform that it will launch a 12-day live event starting tomorrow, showcasing a range of new technologies and innovative projects. Despite ongoing speculation about new models like GPT-5, OpenAI clarified that the successor to GPT-4 will not be released this year.
OpenAI CEO Sam Altman told The Verge that reports about a new AI model named Orion being released in December are "completely fictitious." A company spokesperson also confirmed this but promised that more technological innovations will be released.
Amid expectations for GPT-5, OpenAI may focus more on other new technologies, particularly its latest O1 model. Reports indicate that several X users mentioned the O1 model was temporarily available at "chatgpt.com/?model=o1" before access was blocked. This model supports processing approximately 200,000 tokens and includes image analysis capabilities. Although OpenAI has only released a mini version and a preview of O1, the full version is expected to perform even more robustly in mathematical capabilities.
Additionally, OpenAI's Sora video model has garnered industry attention. This model is set to launch in February 2024 and is continuously improving through testing with selected artists. Although some artists leaked the tool due to lack of compensation, former OpenAI CTO Mira Murati announced that Sora is expected to be officially released in mid-March this year. Now, with other companies also launching similar video generation models, the anticipation for Sora's release has increased.
It's worth noting that despite the release of the GPT-4o model, many anticipated features have yet to be realized, and OpenAI has hardly mentioned these features since the model's release. Now, with the upcoming new technologies like O1 and Sora, OpenAI is once again attracting industry attention, hoping to bring new breakthroughs.