On September 12, OpenAI announced two versions of their new o1 and o1 mini models.
The key feature of these models is that they think about the answer before giving it, just like a human π§
The company estimates that these models can solve physics, chemistry π§ͺ and biology 𧬠problems at the graduate school level, and sometimes even outperform graduate students in math and programming.
Facts: when solving International Math Olympiad problems, GPT-4o answered only 13% of them correctly, while o1 handles 83% of the problems π
β‘οΈ These improved features can be especially useful if you are facing challenging problems in science, programming, math, and related fields.
The o1 can also be used by:
health researchers to annotate cell sequencing data
physicists to create complex mathematical formulas needed for quantum optics
developers in various fields to create and execute multi-step workflows
OpenAI emphasises that for many common tasks ChatGPT is still a great solution, whereas o1 is primarily suitable for really complex tasks in work and study.
π₯ The o1 models are already available in our bot @chat_ai_tg_bot and our site chataibot.pro
