xAI Grok 3 model is 20 times more powerful than Grok 2

AI company xAI has started training the Grok 3 large language model which will be 20 times more powerful than Grok 2 model.

In August this year, xAI announced Grok 2 models for X users. The latest LLM comes with improved response speed and new data processing capabilities.

The model can generate images with a user prompt when you explain it via the Grok interface on the social media site X. You can ask the model to create emails, answer questions, and provide you with recipes. It can also generate code for your application or solve a problem in your snippet.

The company launched the Grok large language model (LLM) in November last year. This chatbot competes against ChatGPT, Google’s Gemini, and other generative AI models.

Earlier this year, xAI started building a supercomputer with Nvidia AI chipset to train future Grok products. The company shared the first images of the cluster in July this year and announced the model in August.

The cluster is developed with enterprise-based solutions with 100,000 Nvidia H100s chipsets. The new supercluster is used to train Grok 2 model and it’s now preparing to do the same with Grok 3.

More Power

Founder Elon Musk said that Grok 3 is training with 10X computing power of the Grok 2. Soon, the firm will improve its computing power to 20 times of its predecessor.

However, Musk’s specific features were revealed through this announcement. However, it’s expected to go online by the end of this year or early 2025.

Meanwhile, Grok 2 is available in two variants Grok 2 and Grok 2 Mini. These models can be accessed via the X app and these two deliver speed and efficiency for their respective users.

(source)