Illon Markas launches Xai Grok 3 model amid Amnesty International's narrow competition


Muhammad Salim Korkutata Anatolia Gety pictures

On Tuesday, Illon Markai revealed the latest model of artificial intelligence, GROK 3, claiming that he could outperform offers from Openai and Chinese Deepseek based on early tests, which included unified tests on mathematics, science and coding.

“We are very excited to present GROK 3, which is a more capable size than Grok 2 in a very short period of time,” Musk said in the Grok 3 demonstration that was broadcast on social media. X.

The team also said it launches a new product called “Deep Search”, which will serve as a “next generation search engine.”

Xai team said Grok 3 will be offered to Premium X subscribers later in the day, and it will also be available through a separate subscription for web versions and applications in the form.

He spoke at the Global Governments Summit in Dubai last week Musk It was called the model “Scary Smart”, with strong thinking capabilities, claiming that it surpassed all other models in the interior tests of Xai.

“This may be the last time that artificial intelligence is better than Grok,” Musk said at the time, adding that he was trained in “a lot of artificial data,” and he was able to think about his mistakes to achieve logical consistency.

The Xai team claimed that the early repetition of GROK 3 has obtained better rating than the current competitors on Chatbot ArenaA group web site stimulates the various Amnesty International models against each other in blind tests.

Near the end of the product offer, Musk said the company will continue to improve the model.

“We must confirm that this is a kind of experimental version, which means that you must expect some defects in the beginning, but we will improve it quickly, almost every day,” he said, adding that the audio aid of the model will be released at a later time.

Intense competition

Musk, Who was very frank About the potential risks of artificial intelligence, Xai began in 2023 when entering the Ontician AI market that includes Chatgpt from Openai.

In September last year, Openai launched its most advanced style, The O1, with which she came Thinking capabilities He was able to solve the tasks of science, coding and mathematics relatively complex.

Musk, alongside Sam Altman, helped create Openai as a non -profit organization in 2015.

However, in recent years of the leadership of Musk and Openai She was stopping. Musk recently led an investor collection presented an offer To buy a non -profit parent for the beginning of AI for $ 97.4 billion – Openai's offer was rejected.

Last month, the Chinese company Deepseek was shocked in the artificial intelligence market when it released an artistic paper claiming that one of the open source models managed to compete with the performance of the O1AI O1 model despite the use of cheaper and less powerful process.

The achievement has been accomplished in the face of the United States, which restricts the leading artificial intelligence chips maker Nafidia From selling advanced graphics processing units-used to train artificial intelligence models-to China.

Xai has “Colossus SuperCanting,“With regard to artificial intelligence training, who said last year he used a group of 100,000 advanced NVIDIA graphics processing units to train artificial intelligence. On Tuesday, the company revealed that it doubled the size of its GPU group to train GROK 3.

While many artificial intelligence and technology experts CNBC said That Deepseek has intensified artificial intelligence competition, which indicates what can be done with less advanced technology, while others are more doubtful of its effect.

Leave a Reply

Your email address will not be published. Required fields are marked *