Financial Associated Press, August 28 (Editor Shi Zhengcheng) Earlier this month, Sam Altman, CEO of the technology start-up OpenAI, posted a photo of his garden on the Internet, showing four small strawberries gently growing The ground hangs on the branches. The latest news on T

Financial Associated Press, August 28 (Editor Shi Zhengcheng) Earlier this month, Sam Altman, CEO of the technology start-up openai, posted a photo of his garden on the Internet, showing four small strawberries. Hanging gently on the branches. The latest news from

on Tuesday showed that the important information hidden in that picture - openai's next blockbuster AI product, a new generation inference model code-named "Strawberry Project" will be available as soon as this fall.

(Source: x)

First of all, I need to explain why the industry is so excited about this "strawberry".

Similar to self-driving cars, OpenAI divides the realization of general artificial intelligence (AGI) into 5 steps: The first is the lowest level we are currently at, conversational AI such as chatgpt.

is the second layer where "Strawberry" is located - inference AI. Large models can handle more complex tasks, have more comprehensive capabilities, and are closer to human proficiency. The following layers are related to the characteristics of autonomy, innovation and organization.

will be available as soon as this fall.

U.S. media quoted people familiar with the project as saying that OpenAI aims to launch the reasoning AI code-named "Strawberry" as soon as this fall. This project was also previously known as q*. The specific product form of

may be part of the chatgpt upgrade. 's new AI will be able to solve "never-before-seen math problems" and word puzzles published in the New York Times. These are tasks that existing AI cannot reliably complete.

In addition to understanding math problems, the "strawberry model" can also be used to program and answer a range of more subjective questions, including handling some multi-step tasks. openai and many competing products hope that product upgrades will bring more revenue opportunities. The latest news revealed that openai can currently earn US$283 million in revenue every month through enterprise services and chatgpt subscriptions, but the company is still in a state of expansion.

It is reported that there are multiple versions of the "Strawberry Model". A reduced and simplified version is used for chatbots, while the full version of the model has a special mission - generates training data for the next generation flagship large model Orion, which solves The problem of shortage of high-quality texts on the Internet.

With the main competitors anthropopic and meta, they are trying to catch up. Now OpenAI also needs an AI with incredible capabilities to reaffirm its status as an "industry leader."

is more accurate and slower.

is one of the few major updates of openai at the end of this year. The company launched gpt-4o, which is known for its speed, earlier this year. However, according to the latest reports, the increase in the reasoning capabilities of 's "Strawberry Model" will come at the expense of 's response speed. Therefore, the main scenarios of this application should be those applications that have little demand for feedback speed.

So it is not difficult to understand that for scenarios such as gpt search that require immediate and quick response, I am afraid that this project is not suitable for use.