Previously, Elon Musk announced that training for Grok 3 was complete and the model possessed superior computing capabilities.
xAI revealed that it used 200,000 Nvidia H100 GPUs at a data center in Memphis to train this model. The company also stated that, along with Google Gemini and ChatGPT 4-o, Grok 3 is considered a competitor to DeepSeek-V3 and Claude 3.5 Sonnet.

The Grok 3 includes several sub-models, including the Grok 3 Reasoning and the Grok 3 mini. The Grok 3 mini is designed for faster response times, although this may come at the cost of some accuracy.
xAI also conducted several tests on Grok and GPT-4o. The results showed that Grok 3 outperformed GPT-4o on several physics and biology questions at a doctoral level. The model has begun deployment and achieved competitive scores in Chatbot Arena – where users can vote for the model they believe is most accurate.
Grok is now available on the App Store in the US, Australia, and India, while Android users still need to access Grok through X.
Source: https://kinhtedothi.vn/phien-moi-nhat-cua-chatbot-ai-grok-duoc-cong-bo.html






Comment (0)