Elon Musk just released Amnesty International than ChatGPT – this is why this is important

Photo of author

By [email protected]


Join daily and weekly newsletters to obtain the latest updates and exclusive content to cover the leading artificial intelligence in the industry. Learn more


ELON Musk AI xi unveil Groc 3The latest model of artificial intelligence that the company claims to surpass competitors through the main technical standards. The advertisement represents a major escalation in the race to develop the most powerful artificial intelligence systems.

The launch comes after only days Musk 97.4 billion dollars failed to get OpenaiThe company that he participated in its founding with SAM Altman in 2015. During a demonstration reserved on X, Musk Grok 3 described as “an arrangement of Grok 2” and confirmed his ability to think through complex problems.

Early test appears to support some Xai claims. It tops the influential model Chatbot Arena Total board, higher record of Openai’s GPT-4OGoogle twin and V3 Deepseek Form In the blind user test. Grok 3 criteria show superior degrees in mathematics (AIME ’24), scientific thinking (GPQA) and coding tasks.

Grok 3 leads Chatbot Arena team with a score of about 1400, greatly outperformed the other AI models in the blind user test. (Source: XAI)

The huge computer infrastructure in GROK 3: 200,000 graphics processing units and a new data center

“Grok 3 clearly has the capabilities of thinking about the latest ideas.” Andrej Karpathy In the X post after the early access test. “A few models get this correctly right. The first Openai thinking models also get, but Deepseek-R1, Gemini 2.0 Flash, and Claude La.”

The development of the model requires huge mathematical resources. Xai has doubled its GPU group to 200,000 NVIDIA chips for training, which is located at the new Memphis Data Center. This investment in infrastructure highlights the increasing mathematical requirements for the development of advanced artificial intelligence, as companies are racing to build more capable systems.

Deep search and advanced thinking: How GROK 3 aims to bypass Chatgpt and Google Gemini

The main innovation is the “Deepsearch” feature for Grok 3, which combines the web search and thinking capabilities to analyze information from multiple sources. The system also includes specialized patterns to solve complex problems, including the “thinking” function that shows the thinking process and the “big brain” that allocates an additional computing power for difficult tasks.

“The thing that you should really pay attention to in AI is the speed of learning. XAI learns a way faster than anything else, Robert ScobleQuoting a conversation with the founder of Apple Siri Tom Gruber.

However, some restrictions appeared during the test. Karpathy indicated that the model sometimes makes categories and fights with certain types of humor and moral tasks. These challenges are common through current artificial intelligence systems and highlighting the continuous difficulties in developing a human -like artificial intelligence.

Scale.ai CEO Alexander Wang Praise the version, “Grok 3 is the best new model in the world from the Xai team!” He referred to her superior performance in various standards and expressed his enthusiasm for future cooperation.

The artificial intelligence industry competition is rising: What the GROK 3 launch for Openai and Deepseek and the future of artificial intelligence means

The model will be available through the X’s Premium+ ($ 40 per month) subscription and with a new independent.SupergrokService ($ 30/month). API has been planned for the Foundation for the coming weeks.

This launch intensifies competition in the artificial intelligence industry, especially as the Chinese operating start Dibsic A similar performance has recently shown with low mathematical requirements. Development also raises questions about the sustainability of the armament race in artificial intelligence, as companies invest billions of dollars in the infrastructure of the increasing devices.

In the main performance standards, Grok 3 and its mini variables appear superior through mathematics, science and coding tests compared to the competing models of Google, Openai, Anthropic and Deepseek. The full Grok 3 (Dark Blue) model has particularly strong results in scientific thinking. (Source: XAI)

Musk confirmed that Grok 3 is still in the experimental version, with improvements to improvements.Almost every day“The company plans to add the capabilities of the audio reaction within weeks and will open its previous source, GROK 2, as soon as the new version stabilizes.

However, the most amazing aspects of GROK 3 may be for the first time their technical specifications or standards, but what it represents: The escalating tension between Musk and his former colleagues in Openai. A few days after his failed failure of $ 97.4 billion to get Openai, Musk revealed a model that challenges his superiority-which indicates that in the high risk race for artificial intelligence dominance, the suitor can become a huge competitor.



https://venturebeat.com/wp-content/uploads/2025/02/nuneybits_Vector_art_of_Elon_Musk_in_front_of_a_very_large_X_7af9382a-4f24-4e89-880e-26f99109458d.webp?w=853?w=1200&strip=all
Source link

Leave a Comment