Elon Musk’s Xai announced the Grok-3 on Tuesday, revealing that the new artificial intelligence model will have “more than 10 times the computational power of its predecessor.” Xai said its latest flagship outperformed Openai’s GPT-4o, Google’s Gemini and Deepseek’s V3 models in early testing, and now has “advanced reasoning” capabilities .
The so-called inference model is intended to provide stronger results, breaking down into smaller tasks into more complex questions, and trying to fact-check yourself before providing solutions, resulting in more complex questions. You are trained to answer. Similar models have been developed by rival companies such as Openai’s O1, Deepseek’s R1, and Google’s Gemini Flash Thinking.
There are two Grok-3 inference modes available. “Think” displays Grok’s inference when resolving a request. “Big brain” for complex tasks that require more computing power. Xai also launches a Grok AI agent product called Deep Search. This is what the company calls the “next-generation search engine.”
GROK-3 inference functionality is available on the GROK app for X Premium Plus subscribers. This was the second Premium Plus hike in two months, up from $16 to $22 in December. Xai also said it will launch a new subscription plan called SuperGrok, which will provide “the quickest access to the most advanced features and new features.” SuperGrok reportedly costs $30 a month, but it’s unclear whether this is an additional charge on top of the X subscription.
Elon Musk said Grok Chatbot will soon acquire a synthesized voice feature similar to Openai’s Advanced Voice mode in ChatGpt. Xai also plans to create open source for Grok-2 in the coming months.
