OpenAI is about to release its latest artificial intelligence product, and this time, it can do more than help you with remedial tasks and answer knowledge-based questions.
According to a report from The Information, the AI developer plans to release “Strawberry,” its reasoning-focused AI model, in the next two weeks.
AI that can reason
The model will be a part of the ChatGPT service but is different from the standard conversational AI that has already taken the world by storm. Strawberry will be able to “think” before responding to questions rather than treating it like an immediate answer to a query.
Two people close to the matter told The Information that “thinking” response times will be about 10 to 20 seconds. Strawberry will initially only be able to take in text and produce text back. It will not immediately be multimodal, meaning that upon its initial release, it will not be able to produce image responses.
Currently, OpenAI’s most advanced model available to the public, GPT-4o, provides users with GPT-4-level intelligence but with fast response times and abilities across text, voice and vision responses.
While not much information has been made public about the details of the Strawberry release, OpenAI’s co-founder and CEO, Sam Altman, has been teasing the product for some months.
In July, reports began to surface about OpenAI’s new model, codenamed Strawberry. Initially, it was expected to be announced in November 2023, but its release has come two weeks earlier than the rumored timeline.
In the reports, it was said that Strawberry would effectively scan the internet on its own and conduct “deep research,” as described by OpenAI. This will allow the model to solve more complex, real-world problems at scale.
Then, last month, Altman caused a stir online after he posted a photo of ripe-looking strawberries with the caption, “I love summer in the garden.” Most X users speculated that it was a hint at the forthcoming model.
Related: AI deepfake attacks will extend beyond videos and audio — Security firms
OpenAI on the rise
Since releasing ChatGPT in November 2022, OpenAI has gained significant attention and influence in the tech industry.
On Sept. 5, the San Francisco-based AI company said that paid users have grown 67% since April to break through the 1 million benchmark across its business services, including ChatGPT Enterprise, Team and Edu.
User numbers previously rested around the 600,000 mark in April.
The company also allegedly plans to release more expensive subscription plans for its upcoming large language models, such as the Strawberry and Orion AI models — some of which have been rumored to cost upward of $2,000 per month.
AI Eye: AI drone ‘hellscape’ plan for Taiwan, LLMs too dumb to destroy humanity