OpenAI who is already leading the AI market with ChatGPT is expanding its capabilities to dominate further. To do so, OpenAI is looking to include a new AI model which has a mind of its own and carries a capability to think before answering a query. OpenAI is calling it reasoning-focused artificial intelligence and this ability to question and make a decision enables this AI model to solve more complex problems. This also improves the quality of the output provided by ChatGPT. Its ability to think before giving an answer greatly improves the quality and authenticity of the results.

OpenAI plans to include Strawberry within current version of the ChatGPT within the next 2 weeks time.

Let’s dive deeper to understand Strawberry AI a little more.

What is Strawberry AI?

Strawberry AI is OpenAI’s yet to launch AI model that is designed to solve complex problems with the ability to think and analyze real time queries. Rather then providing instant answer to user queries, Strawberry AI generates more considered results. This allows the model to give more appropriate answers and avoid hallucinations. As per reports, Strawberry AI can only respond to text queries with text results unlike the current AI models that also have the ability to read images. Which means it is not multimodal yet, but gives better capabilities to solve more complex problems including solving advanced algebraic equations and handle handle intricate and nuanced enquiries.

The response time of Strawberry AI as reported to range anywhere between 10 seconds to 20 seconds for each query, which is notably longer. This however is due to the time it takes to analyze the problem and make decisions in real-time. For generic queries that do not involve complex reasoning methodologies can be answered instantly. OpenAI may have to restrict usage of this model with different pricing tiers and limit queries submitted in a given timeframe.

What else is happening in OpenAI?

OpenAI continues to work on new AI advancements like Strawberry Orion and GPT-NEXT. Recently, during the KDDI Summit 2024, OpenAI Japan CEO, Tadao Nagasaki disclosed GPT-Next is another product set to launch soon. GPT Next claims to be 100 time better then the current GPT-4.

According to Nagasaki, the GPT-Next model will incorporate a scaled-down version of OpenAI’s confidential Project Strawberry, previously referred to as Project Q* or Q-star. GPT-Next is expected to deliver enhanced performance, not solely from greater computational power, but due to advancements in its core architecture, which enable more efficient and powerful processing capabilities.