The Launch of Grok 3 and Its Purported Capabilities
In a highly anticipated livestream event, Elon Musk’s artificial intelligence company, xAI, has launched its latest flagship AI model, Grok 3. According to Musk, the Grok 3 models are “an order of magnitude more capable than Grok 2.” The Grok 3 family consists of three main models: the Grok 3 mini, Grok 3 Reasoning, and Grok 3 mini Reasoning.
While the Grok 3 mini is designed for quick question-answering, it lacks the accuracy of its more advanced counterparts. The Grok 3 Reasoning and Grok 3 mini Reasoning models, however, are claimed to possess human-like reasoning capabilities. These models are touted as being able to mimic the way humans analyze and process information, potentially revolutionizing how we interact with AI assistants.
Grok 3’s Claimed Superiority and Pricing
xAI confidently claimed during the event that the Grok 3 Reasoning model outperformed OpenAI’s highly regarded o3-mini model on several benchmarks, according to TechCrunch. This bold assertion positions Grok 3 as a potential frontrunner in the race for advanced AI reasoning capabilities.
Initially, the Grok 3 features will be available to subscribers of X’s Premium+ tier, which now costs a staggering $40 per month in the United States. This represents a significant price hike from the previous $16 monthly fee, which was increased to $22 in December 2024, as reported by Engadget. Alternatively, users can opt for a separate subscription option called SuperGrok, expected to cost $30 per month, according to leaked information.
Grok 3’s Advanced Features and Capabilities
With the Grok 3 models enabled, users will have access to advanced reasoning capabilities by using the “Think” command for mathematics, science, and programming queries. For even more complex inquiries, the “Big Brain” function can be invoked, which requires additional computing power.
One of the standout features of Grok 3 is DeepSearch, which xAI describes as the “next generation search engine.” DeepSearch will scour the internet and X (formerly Twitter) to provide concise summaries for research inquiries, leveraging the models’ reasoning capabilities.
Additionally, xAI announced that the Grok app will soon receive a “voice mode” update, introducing synthesized voices for conversational interactions with users. Furthermore, the company plans to open-source its older Grok 2 models in the coming months.
As the AI race intensifies, xAI’s Grok 3 launch represents a bold step forward in the pursuit of human-like reasoning capabilities in artificial intelligence. However, only time will tell if the claimed superiority of Grok 3 over its competitors holds true in real-world applications.
This article originally appeared on Engadget.