OpenAI is planning to roll out a new artificial intelligence model, named ‘Strawberry,' aimed at enhancing chatbots' reasoning abilities. According to a report from The Information, the model will be released during October. The AI focuses on tackling intricate tasks such as advanced mathematics and programming, areas where current systems have limitations.
Rebranding and Technical Capabilities
Originally named Q*, the model has been renamed ‘Strawberry.' The new AI is set to enhance OpenAI's present systems considerably. Sources told The Information that OpenAI presented ‘Strawberry' to U.S. national security officials during the summer, pointing to its potential applications in that field.
Earlier this month, OpenAI CEO Sam Altman hinted at the new AI model by sharing an image of strawberries. The cryptic message became clear with the official unveiling of ‘Strawberry.' OpenAI's model is expected to bolster chatbot capabilities, allowing them to undertake tasks that require a higher level of reasoning and problem-solving.
Strategic Importance and Future Prospects
Introducing ‘Strawberry' is part of OpenAI's strategy to advance conversational AI. By addressing the shortcomings of current models, ‘Strawberry' aims to improve chatbot reliability and functionality across various applications. The move underscores OpenAI's commitment to remaining competitive in the ever-evolving tech sector.
OpenAI is also working on another AI model named ‘Orion,' which aims to exceed GPT-4's capabilities, with ‘Strawberry' playing a significant role. Insiders indicate that a chatbot version of ‘Strawberry' might debut this fall, potentially integrated into ChatGPT. The model is designed to solve complex math problems and optimize programming tasks, with enhanced logic that improves language-related challenge-solving.
Internal Testing and Performance
In internal trials, ‘Strawberry' demonstrated its prowess by solving the New York Times word puzzle “Connections.” An AI tested internally by OpenAI, likely ‘Strawberry,' scored over 90 percent on the MATH benchmark. Internal documents reveal plans to use ‘Strawberry' for autonomous internet searches, enabling it to conduct comprehensive research and planning.
The development approach for ‘Strawberry' is similar to Stanford's “Self-Taught Reasoner” (STaR) method, designed to boost AI reasoning skills. Former OpenAI chief researcher Ilya Sutskever, now running a startup focused on secure super AI, is credited with providing the foundational ideas for ‘Strawberry.'
Google DeepMind is also delving into AI systems with advanced mathematical abilities, having created AlphaProof and AlphaGeometry 2, which achieved silver at the International Mathematical Olympiad. However, it remains unclear how scalable and generalizable these models will be.
