Elon Musk's xAI on Monday released the latest model of its artificial intelligence (AI), Grok 3, claiming it surpasses competitors in math, science, and coding. It also introduced new features like DeepSearch and voice mode for premium users.
Grok 3 offers advanced reasoning capabilities for complex queries through its "Think" and "Big Brain" modes. It can be accessed through X's Premium+ at $40/month or SuperGrok at $30/month or $300/year. The new monthly price is up almost 82% from its previous price of $22.
The development of Grok 3, which Musk spoke about at the recent World Governments Summit in Dubai, was accelerated by xAI's Colossus supercomputer, utilizing 100K Nvidia H100 GPUs and 200M GPU hours for training, which is 10 times more than its predecessor Grok 2.
If Musk's description of Grok 3 is accurate, this new chatbot will become the world's smartest AI model. It's not just three times faster than the remarkable Grok 2, but it will offer assistance in multiple languages and use customer feedback to enhance accuracy. The key to a successful AI chatbot is its ability to train itself and Grok 3 is proving very capable.
Grok 3 will likely be on par with, if not slightly inferior to, its ChatGPT counterparts, as stated by one of xAI's engineers himself. The only thing Grok has that other AI companies lack is its ability to scrape real-time data from X posts, but that's made more difficult by misinformation. Grok has a way to go before it can be crowned king of AI.
No matter how tech executives promote their products, every new AI development is part of an arms race to human-technology singularity. These rapidly advancing models, especially in the hands of governments, will undoubtedly be used by humans for military purposes, and eventually by the AI bots themselves as a once-fictional Terminator becomes reality.