Today, OpenAI launched the OpenAI o1 language model, which “thinks” before answering questions. According to the official demo, OpenAI o1 processes the question first, and when it responds, it sends both its thoughts and the answer together. Although only ChatGPT Plus and team users can access it for now, we can still get a preview of what OpenAI o1 will be able to do in the future. Today, I’ve compiled the confirmed capabilities of OpenAI o1.
ChatGPT’s New Model: OpenAI o1’s Logical Reasoning and Use Cases
Following GPT-4o, which can think and react instantly, today OpenAI launched “OpenAI o1,” a model that contrasts with GPT-4o. OpenAI o1 emphasizes thinking through a problem before responding, providing users with more contextual and logical reasoning in its replies. For those interested in learning more about the details of OpenAI o1, you canGo read the related report.Today, we will mainly introduce the practical applications of OpenAI o1.
According to the practical application of OpenAI o1 provided by OpenAI, because OpenAI o1 is able to think, it can solve complex logic puzzles. For example, given the question “A princess is as old as the prince will be when the princess is twice as old as the prince was when the princess’s age was half the sum of their present age.”, asking how old the princess and the prince are now? OpenAI o1 can work out the appropriate answer through a series of complex calculations:
OpenAI o1 solves a complex logic puzzle. pic.twitter.com/rpJbh8FkAg
— OpenAI (@OpenAI) September 12, 2024
OpenAI o1 codes a video game from a prompt. pic.twitter.com/aBEcehP0j8
— OpenAI (@OpenAI) September 12, 2024
OpenAI o1 is capable of thinking, so it can effortlessly answer questions that are extremely tricky for other large language models. Ask OpenAI o1 how many r’s are in Strawberry, and after a few seconds of thought, it can give the correct answer. This greatly reduces the hallucinations or rambling that were common in earlier large language models.
OpenAI o1 answers a famously tricky question for large language models. pic.twitter.com/5ZlQIOBWEd
— OpenAI (@OpenAI) September 12, 2024
Besides programming, OpenAI o1 has also made considerable progress in mathematics. OpenAI o1 can now solve Nonogram puzzles:
Besides the aforementioned, OpenAI o1 can actually do many other things now—its professional knowledge responses in fields like genetics, quantum physics, and economics have also been highly praised. OpenAI has officially stated that o1 can reason through complex tasks and solve harder problems in science, coding, math, and more that previous models couldn’t answer. Although only ChatGPT Plus and Team users can access OpenAI o1 for now, OpenAI says enterprise and education users will also get it a week later. With a little patience, we’ll be able to experience ChatGPT’s latest thinking models, o1 and o1 mini, ourselves.
However, everyone shouldn’t have too high expectations for OpenAI o1 just yet, because after testing by netizens, it still gives wrong answers to very basic questions.
这就是你们说的推理能力强了？博士级别？🤨 pic.twitter.com/XwonwB79P5
— 小互 (@imxiaohu) September 13, 2024
If you’d like to learn more about OpenAI o1, you can click the link below to read the related report:
會思考（Thinking）的 ChatGPT 最新模型 o1 與 o1 mini 來了，志在陪伴研究者脫離難題的「兔子洞」
Source: KOCPC Chinese