OpenAI has been working on a Strawberry AI engine that aims to significantly improve reasoning capability and solve complex math and coding problems. Last month, we reported that OpenAI was planning to integrate Strawberry into ChatGPT as early as this fall. Now, a new report by The Information reveals that OpenAI has changed the release timeline and it could launch in the next two weeks.
While it’s not confirmed, OpenAI might announce Strawberry at its upcoming DevDay 2024 event. The report quotes two people who tested Strawberry and shared their insights. Strawberry takes 10 to 20 seconds to “think” before generating an output. It takes the extra time to process complex information and respond to queries accurately.
That said, those who tested Strawberry say that waiting 10 to 20 seconds for a slightly improved answer than GPT-4o doesn’t seem worth it. In addition, Strawberry takes the extra seconds even for simple queries, which adds to the latency. Moreover, Strawberry fails to adhere to users’ coding preferences.
Strawberry, of course, uses more compute resources and generates more tokens before each response. In all likelihood, OpenAI might hike the subscription fee of ChatGPT Plus. The Information earlier reported that OpenAI is considering hiking the subscription fee to as high as $2000 per month. It sparked a debate over whether individuals need to pay a massive sum to access intelligent AI systems in the future.
What You Will Learn
- The anticipated launch timeline and capabilities of OpenAI's Strawberry AI engine.
- Insights from early testers about the performance and latency of Strawberry AI.
- Potential subscription fee increases and their implications for users.
- The ongoing discussions regarding accessibility and affordability of advanced AI technologies.
