OpenAI may launch its new AI ‘Strawberry’ in two weeks – report

OpenAI is planning to release its new reason-focused AI, “Strawberry,” in two weeks, according to reports. The Information cited two people who reportedly tested the model, stating that the ChatGPT service may be released earlier than its original fall timeline.

While Strawberry is part of ChatGPT, it is expected to be a standalone service. ReadWrite reported on the model set, which is said to have improved AI reasoning skills, enabling it to conduct independent research on the Internet. This would allow the AI to respond to user queries, as well as “think” and gather information independently. The thinking stage is said to last around 10 to 20 seconds.

Strawberry, believed to be an upgrade of the company’s earlier Q* project, is expected to be a major improvement in AI development. Q* had stirred internal debates and was linked to the temporary dismissal of CEO Sam Altman in November 2023, according to insiders.

However, while testers found its performance slightly better than GPT-4o, Strawberry was reported to struggle with short, simple queries and has issues with memory integration. The model lacks image integration, making it exclusively text-based for now. It is expected that Strawberry will also have rate limits and might introduce a higher-priced tier for users seeking faster response times, diverging from the current pricing structure of ChatGPT.

When previously approached about the project and its details, an OpenAI spokesperson had stated, “We want our AI models to see and understand the world more like we do. Continuous research into new AI capabilities is a common practice in the industry, with a shared belief that these systems will improve in reasoning over time.”

OpenAI has been working on several projects in addition to Strawberry including Orion, and others. Earlier this month, Tadao Nagasaki, the lead executive for OpenAI’s operations in Japan, teased an upcoming advancement in large language models (LLMs) from the company, which he referred to as “GPT Next.”

OpenAI reveals GPT Next

OpenAI Japan's representative, Tadao Nagasaki, spoke at KDDI's business event "KDDI SUMMIT 2024" on September 3, 2024.

Key points from his presentation include:

ChatGPT has surpassed 200 million active users as of late August, making it the fastest… pic.twitter.com/av89t5y3La

— Carlos E. Perez (@IntuitMachine) September 3, 2024

Nagasaki said: “The AI model called GPT Next, which will eventually come out, will evolve nearly 100 times based on past achievements.” The image displayed on the monitor behind Nagasaki showed his slide deck, which included a timeline of GPTs. It began with the “GPT-3 Era” in 2021, followed by the “GPT-4 Era” in 2023, then the “GPT Next” era slated for 2024, and concluded with “Future Models” marked as “202x”.

ReadWrite has reached out to OpenAI for comment.

Featured image: Ideogram

The post OpenAI may launch its new AI ‘Strawberry’ in two weeks – report appeared first on ReadWrite.