OpenAI is reportedly working on a new artificial intelligence (AI) model that could significantly improve the math and reasoning capabilities of its own chatbot, ChatGPT. The alleged AI model has reportedly been in the works for several months, but the company has kept it under wraps. The existence of the large language model (LLM) was first reported in July. The Strawberry AI model is reportedly being used to train the LLM, which OpenAI sees as the successor to GPT-4.
OpenAI is reportedly set to launch a Strawberry AI model soon
The information reports that the ChatGPT maker’s new AI model will be launched in the fall. However, there are uncertainties about how OpenAI could introduce the AI model. One possibility, according to the publication, is that the model could be launched as a new chatbot, while the other possibility is that its capabilities could be added to the existing ChatGPT.
The reason Strawberry has received so much attention is reportedly due to its advanced capabilities to solve mathematical and reasoning problems. It is said that the AI model can solve problems that it has not seen before.
This is important because AI models are essentially deep learning models that use algorithms to understand and process different patterns. However, when these patterns and logic become symbolic or vague, the AI struggles. As a result, modern AI chatbots are unable to solve complex mathematical or reasoning problems, or problems that involve contextual knowledge or ambiguity.
If the reports are to be believed, Strawberry has shown that it can outperform current AI models. Not much is known about this AI model in terms of architecture or parameters. However, according to the report, this AI model was previously known as Q* (pronounced Q-star).
Interestingly, after the secretive Strawberry AI model was first reported in a Reuters report, Altman took to X (formerly known as Twitter) to share a cryptic after about strawberries, probably to acknowledge their existence.
