TL; DR
- According to xAI, the company used its Colossus supercomputer to train Grok 3, along with 100,000 Nvidia H100 graphics processors.
- In order to address privacy issues related to training its AI, Grok 3 has used synthetic datasets.
- Further, it also includes self-correction mechanisms, which help the AI identify and correct the mistakes it makes while responding to users.
On Monday, xAI unveiled its latest AI model, Grok 3. With real-time information and multimodal capabilities, the new chatbot is designed to provide insightful solutions to complicated problems. It competes with popular chatbots such as OpenAI’s ChatGPT and Google’s Gemini. Here’s everything we know about the new AI model.
Also Read: The New Dimensity 6400 SoC Shares A Lot Of Similarities With Its Predecessor
What Musk Says About Grok 3
“Grok 3 has very powerful reasoning capabilities, so in the tests that we’ve done thus far, Grok 3 is outperforming anything that’s been released, that we’re aware of, so that’s a good sign,” says Musk in a public statement. “At times, I think Grok 3 is scary smart,” the tech billionaire believes.
In addition to praising his latest ChatGPT, Gemini, and DeepSeek competitor, Musk also announced a separate subscription called Super Grok for all the enthusiasts and fans who want the most advanced capabilities. However, the exact pricing and perks of the subscription model aren’t clear yet.
xAI is also working on a voice interaction feature like Gemini Live. “The goal is to make it so you can talk to it just like you would a person,” says Musk.
Also Read: Tesla Starts Hiring For Roles In Mumbai And Delhi, Might Enter India Soon
How Does Grok 3 Compare To Grok 2?
According to xAI, the company used its Colossus supercomputer to train Grok 3, along with 100,000 Nvidia H100 graphics processors. Together, the GPUs provided a whopping 200 million GPU-hours for training the chatbot. The company also claims that this is 10 times the hours that went into training the Grok 2 chatbot.
In order to address privacy issues related to training its AI, Grok 3 has used synthetic datasets. As a result, Grok 3 can interpret and analyze larger datasets in less time, increasing learning efficiency. The Musk-led company has also included a reinforced learning model, wherein the AI model learns from rewards/penalties based on correct/incorrect responses.
Further, it also includes self-correction mechanisms, which help the AI identify and correct the mistakes it makes while responding to users. It is a simple yet clever technology that allows the AI to compare its responses with those known to be correct and improve its accuracy over time.
Also Read: [Exclusive] Nothing Phone (3a) Pro to Feature a 3X Lens with 60X Hybrid Zoom
Grok 3 vs. Other Popular Chatbots
According to X user @MarioNawfal, Grok 3 is the first-ever AI model to score over 1400 points in Arena, a benchmark platform for large language models that puts up chatbots against each other and relies on human votes for deciding which one is the best.
Who Can Use Grok 3 AI Language Model?
For now, Grok 3 is accessible to Premium Plus subscribers on X, which now costs Rs. 1,750 per month in India or Rs. 18,300 per year (up from Rs. 1,300 and Rs. 13,600 respectively, after the recent price hike in December 2025).
You can follow Smartprix on Twitter, Facebook, Instagram, and Google News. Visit smartprix.com for the latest tech and auto news, reviews, and guides.
