Skip to main content

OpenAI’s advanced ‘Project Strawberry’ model has finally arrived

chatGPT on a phone on an encyclopedia
Shantanu Kumar / Pexels

After months of speculation and anticipation, OpenAI has released the production version of its advanced reasoning model, Project Strawberry, which has been renamed “o1.” It is joined by a “mini” version (just as GPT-4o was) that will offer faster and more responsive interactions at the expense of leveraging a larger knowledge base.

It appears that o1 offers a mixed bag of technical advancements. It’s the first in OpenAI’s line of reasoning models designed to use humanlike deduction to answer complex questions on subjects — including science, coding, and math — faster than humans can.

Recommended Videos

For example, during testing, o1 was fed a qualifying exam for the International Mathematics Olympiad. While its predecessor, GPT-4o, only managed to correctly solve 13% of the problems presented, o1 got 83% of them right. In an online Codeforces competition, o1 scored in the 89th percentile. What’s more, o1 can respond to queries that stumped previous models (like, “which is bigger, 9.11 or 9.9?”). However, the company makes clear that this release is only a preview of the neophyte model’s full capabilities.

The new o1 “has been trained using a completely new optimization algorithm and a new training dataset specifically tailored for it,” OpenAI’s research lead, Jerry Tworek, told The Verge. Using a combination of reinforcement learning and “chain of thought” reasoning, o1 reportedly returns more accurate inferences than its predecessor. “We have noticed that this model hallucinates less,” Tworek said, however, “we can’t say we solved hallucinations.”

Both ChatGPT-Plus and Teams subscribers will be able to test out o1 and o1-mini beginning today. Enterprise and Edu subscribers should have access by next week.

The company says that o1-mini will eventually become available to free-tier users, though it did not specify a timeline. Developers will notice a steep increase in the API pricing for o1, compared to GPT-4o. Access to o1 will cost $15 per million input tokens (compared to $5 per million for GPT-4o) and $60 per million output tokens, four times more than 4o’s $5 per million fee. The real question is whether the new model thinks the word “strawberry” contains two R’s or three.

Andrew Tarantola
Andrew Tarantola is a journalist with more than a decade reporting on emerging technologies ranging from robotics and machine…
Microsoft Copilot: how to use this powerful AI assistant
Man using Windows Copilot PC to work

In the rapidly evolving landscape of artificial intelligence, Microsoft's Copilot AI assistant is a powerful tool designed to streamline and enhance your professional productivity. Whether you're new to AI or a seasoned pro, this guide will help you through the essentials of Copilot, from understanding what it is and how to sign up, to mastering the art of effective prompts and creating stunning images.

Additionally, you'll learn how to manage your Copilot account to ensure a seamless and efficient user experience. Dive in to unlock the full potential of Microsoft's Copilot and transform the way you work.
What is Microsoft Copilot?
Copilot is Microsoft's flagship AI assistant, an advanced large language model. It's available on the web, through iOS, and Android mobile apps as well as capable of integrating with apps across the company's 365 app suite, including Word, Excel, PowerPoint, and Outlook. The AI launched in February 2023 as a replacement for the retired Cortana, Microsoft's previous digital assistant. It was initially branded as Bing Chat and offered as a built-in feature for Bing and the Edge browser. It was officially rebranded as Copilot in September 2023 and integrated into Windows 11 through a patch in December of that same year.

Read more
OpenAI uses its own models to fight election interference
chatGPT on a phone on an encyclopedia

OpenAI, the brains behind the popular ChatGPT generative AI solution, released a report saying it blocked more than 20 operations and dishonest networks worldwide in 2024 so far. The operations differed in objective, scale, and focus, and were used to create malware and write fake media accounts, fake bios, and website articles.

OpenAI confirms it has analyzed the activities it has stopped and provided key insights from its analysis. "Threat actors continue to evolve and experiment with our models, but we have not seen evidence of this leading to meaningful breakthroughs in their ability to create substantially new malware or build viral audiences," the report says.

Read more
From Open AI to hacked smart glasses, here are the 5 biggest AI headlines this week
Ray-Ban Meta smart glasses in Headline style are worn by a model.

We officially transitioned into Spooky Season this week and, between OpenAI's $6.6 million funding round, Nvidia's surprise LLM, and some privacy-invading Meta Smart Glasses, we saw a scary number of developments in the AI space. Here are five of the biggest announcements.
OpenAI secures $6.6 billion in latest funding round

Sam Altman's charmed existence continues apace with news this week that OpenAI has secured an additional $6.6 billion in investment as part of its most recent funding round. Existing investors like Microsoft and Khosla Ventures were joined by newcomers SoftBank and Nvidia. The AI company is now valued at a whopping $157 billion, making it one of the wealthiest private enterprises on Earth.

Read more