The New OpenAI Model that Thinks Before it Answers
By Vukan Ljubojevic | TH3FUS3 Senior Writer
September 15, 2024 11:40 AM
Reading time: 1 minute, 38 seconds
TL;DR OpenAI has unveiled its new AI model, OpenAI o1, which promises advanced reasoning capabilities. The model excels in STEM and coding tasks and aims to revolutionize AI's problem-solving abilities. OpenAI plans to release upgraded versions soon.
OpenAI has released several new artificial intelligence models under a revised naming scheme, starting with its latest OpenAI o1 model. The company claims this model can 'think before it answers,' marking a significant leap in AI capabilities.
A New Benchmark in AI
In a September 12 blog post, OpenAI described the OpenAI o1 model as a groundbreaking advancement in complex reasoning tasks. 'For complex reasoning tasks, this is a significant advancement and represents a new level of AI capability,' the company stated. Given the model's prowess, OpenAI is resetting the counter and naming this series OpenAI o1.
The new models, including OpenAI o1, are designed to take their time to think and utilize 'chain-of-thought' reasoning.
This approach effectively solves complex tasks in STEM (science, technology, engineering, and math) and coding-related problems. OpenAI shared videos showcasing OpenAI o1 coding a video game from a prompt and solving intricate logical puzzles.
Availability and Performance
The OpenAI o1 'preview' and 'mini' models are already available to ChatGPT Plus subscribers. The company plans to release improved versions in the coming months.
OpenAI also shared data indicating that OpenAI o1 outperforms GPT-4o in several benchmarks, including Ph.D.-level science topics in Biology, Chemistry, Physics, and some U.S. high school exams.
"Due to OpenAI o1 mini's focus on STEM reasoning capabilities, it isn't as knowledgeable in other areas outside of its narrow focus," OpenAI said.
'Its factual knowledge on non-STEM topics such as dates, biographies, and trivia is comparable to small LLMs such as GPT-4o mini.'
Future Improvements
OpenAI acknowledged the current limitations of the OpenAI o1 model and promised future enhancements. 'We will improve these limitations in future versions and experiment with extending the model to other modalities and specialties outside of STEM,' the company added.
Industry pundits had anticipated that OpenAI would release a reasoning-focused AI model under the codename Strawberry in September. However, OpenAI does not disclose distinctions between different models under development.
