The OpenAI company launched this Thursday the successor of ChatGPT. It is about o1, secretly known as “Strawberry”, a new one generative artificial intelligence (AI) capable of reason and of answer particularly complex questionsfor example in mathematics.
Unlike their predecessors, these new models have been designed to refine your thought processestry different methods and acknowledge errors before providing a final answer.
The CEO of OpenAI, Sam Altman, He praised the models as “a new paradigm: an AI that can perform complex general-purpose reasoning.” However, he cautioned that the technology “It still has flaws, it is limited and it looks more impressive the first time you use it than after spending more time with it.”
OpenAI launches o1, a new artificial intelligence that surpasses ChatGPT
OpenAI, backed by Microsoft, He said that in trials, the models performed comparably to PhD students on difficult physics, chemistry and biology tasks. They also excelled in math and coding, achieving a 83% success rate in a qualifying exam for the International Mathematical Olympiadcompared to 13% for GPT-4o, its most advanced general-purpose model.
In a math contest for American high school students, o1 was “in the top 500”Altman’s company added.
“Just like a human being who can think for a long time before answering a difficult questiono1 uses a chain of thought (…) Learns to recognize and correct his mistakes. Learn to break down the trickier steps into simpler ones. Learn to try a different approach when the current one doesn’t work,” OpenAI explained.
The scope of OpenAI’s new artificial intelligence
The company said that the Enhanced reasoning capabilities could be used by healthcare researchers to annotate cell sequencing datathat the physicists can generate complex formulas and the computer developers are able to build and execute multi-step designs.
He also said that the new models are better able to resist attempts to bypass security mechanisms than previous ones.
The company It noted that its enhanced security measures included recent agreements with the US and UK AI Safety Institutes, which were granted early access to the models for evaluation.
Finally OpenAI released “Strawberry” under the name o1
In mid-July it became known that OpenAI was working on a new artificial intelligence, secretly known as “Strawberry”, according to a copy of a recent internal document from OpenAI seen by Reuters in May. The new model finally came to light on Thursday under the name o1.
At the time, the source described the plan to Reuters as a work in progress. The news agency was unable to determine how close Strawberry was to reaching the public, given that its development was a closely guarded secret by the company.
The document described a project using Strawberry models aimed at enabling the company’s AI to not only generate responses to queries but plan far enough in advance to autonomously and reliably navigate the internet to perform what the company calls “deep research,” the source said.
OpenAI advances in the development of artificial intelligence
OpenAI continues to make its way into the Artificial Intelligence market, generating a strong interest and triggered investments in your company. Thanks to this, since the company confirmed, last week, that it has more than 1 million paying users on their commercial products.
The company was founded in 2015, and Altman was one of the original members who helped found the company, initially starting out as a non-profit. In 2019, he took over as CEO and, after a brief hiatus, remained in that role to revolutionize how artificial intelligence works.
Altman says that to have a balanced and productive lifestyle, it is necessary to focus on three key points: Recognize the important things and put your focus and energy there, do not waste time on irrelevant issues, and make lots of lists.
Source: Ambito
