OpenAI is planning to launch a new AI product in the fall, “Strawberry,” previously known as Q* (pronounced “Q Star”).
This advanced AI model will have capabilities that extend beyond traditional training, allowing it to solve math problems without prior specific training and perform complex tasks like designing market strategies, solving intricate word puzzles, and conducting “deep research.”
According to a report by The Information, Strawberry might be integrated with the latest version of ChatGPT, known as ChatGPT-4o.
Reuters first reported on Strawberry’s development in July 2024, citing an internal document from OpenAI.
This document suggested that Strawberry would generate answers to queries and autonomously navigate the internet to perform complex research tasks.
The AI is designed to think ahead, a significant leap from current models, and aims to address a critical issue in AI—hallucination, where AI generates inaccurate or nonsensical information.
Strawberry is OpenAI’s solution to the ongoing challenge of AI hallucination. Traditional AI models often struggle with common sense problems and logical reasoning that are intuitive to humans.
Strawberry is expected to enhance AI’s ability to recognize logical fallacies, solve intuitive puzzles, and engage in tasks that require forward-thinking and planning.
OpenAI CEO Sam Altman hinted at the upcoming release with a subtle post on August 7, 2024. He shared an image of strawberries on his social media handle with the caption, “i love summer in the garden,” fueling speculation about the product.
Strawberry’s development is part of OpenAI’s broader strategy, which includes creating its next large language model (LLM), codenamed Orion.
This new tool will not only push the boundaries of what AI can achieve but also plays a crucial role as OpenAI seeks to raise additional capital from investors.