OpenAI, the creator of ChatGPT, is developing advanced reasoning technology under the code name 'Strawberry'. This project aims to significantly enhance the reasoning abilities of its artificial intelligence models, allowing them to tackle complex science and math problems and conduct independent online research.
According to a Reuters report released on Friday, OpenAI is working on this advanced technology to improve its large language models (LLMs). The company, led by Sam Altman, is optimistic that Strawberry will mark a substantial leap forward in AI capabilities.
Previously known as Q*, Strawberry is a highly confidential initiative within OpenAI. Early demonstrations of Q* have showcased its potential, with AI models displaying the ability to solve challenging science and math questions that are beyond the reach of current commercial models.
The project aims to enable AI to perform not only advanced problem-solving but also to undertake what OpenAI refers to as "deep research." This involves planning and executing tasks autonomously over extended periods, enhancing the AI’s capability to conduct thorough and reliable internet-based research.
Strawberry will introduce a new method of processing AI models after their initial training on extensive datasets. This post-training approach is designed to refine and adapt the models, improving their performance for specific tasks even after they have been trained on general data.
OpenAI envisions Strawberry as a tool for long-horizon tasks (LHT), which require AI models to strategize and execute a series of actions over an extended timeframe. This advancement is expected to broaden the scope of what AI can achieve, pushing the boundaries of current technological capabilities.
Elon Musk Unveils Grok AI 2 Following Grok AI 1.5, Teases Grok AI 3
PLI Scheme Propels Telecom Equipment Sales Beyond Rs 50,000 Cr
