Elon Musk's AI company, xAI, has officially released its latest flagship AI model, Grok 3, along with significant updates to the Grok apps for iOS and the web. This ambitious launch marks a major milestone for xAI, as Grok 3 promises to surpass its predecessors in terms of capability and performance.

Grok 3 is xAI's answer to models like OpenAI's GPT-4o and Google's Gemini, and it can analyze images and respond to questions. The new model is powered by an enormous data center in Memphis, containing around 200,000 GPUs, and has been trained on an expanded dataset that includes filings from court cases. According to Musk, Grok 3 was developed with "10x" more computing power than its predecessor, Grok 2.

Musk described Grok 3 as a "maximally truth-seeking AI, even if that truth is sometimes at odds with what is politically correct." This emphasis on truth-seeking is a key aspect of Grok 3, which xAI claims beats GPT-4o on benchmarks like AIME and GPQA. The model's performance has also been competitively tested in Chatbot Arena, a crowdsourced platform that pits different AI models against each other.

Grok 3 is not a single model, but rather a family of models, including a smaller version called Grok 3 mini, which responds to questions more quickly at the cost of some accuracy. Two variations, Grok 3 Reasoning and Grok 3 mini Reasoning, can "think through" problems, similar to "reasoning" models like OpenAI's o3-mini. These models thoroughly fact-check themselves before giving out results, which helps them avoid pitfalls that normally trip up models.

xAI claims that Grok 3 Reasoning surpasses the best version of o3-mini on several popular benchmarks, including the AIME 2025 mathematics benchmark. The reasoning models can be accessed via the Grok app, where users can ask Grok 3 to "Think" or leverage "Big Brain" mode for more difficult queries. xAI describes the reasoning models as best suited for mathematics-, science-, and programming-related questions.

In an effort to prevent distillation, a method used by AI model developers to extract knowledge from another model, Musk said that some of the reasoning models' "thoughts" are obscured in the Grok app. This move comes amidst concerns about AI model distillation, as seen in the recent controversy surrounding DeepSeek's alleged use of OpenAI's models.

The Grok app will also gain a new feature called DeepSearch, which scans the internet and X to analyze information and deliver an abstract in response to a question. Subscribers to X's Premium+ tier will get Grok 3 first, while other features are gated behind a new plan called SuperGrok, priced at $30 per month or $300 per year. SuperGrok unlocks additional reasoning and DeepSearch queries, as well as unlimited image generation.

In the future, the Grok app will gain a "voice mode," giving Grok models a synthesized voice. The Grok 3 models will also arrive in xAI's enterprise API, along with the DeepSearch feature. Furthermore, xAI plans to open-source Grok 2 in a few months, following its general approach of open-sourcing the last version of Grok when the next version is fully released.

When Musk first announced Grok, he pitched the AI as edgy, unfiltered, and anti-"woke" – willing to answer controversial questions other AI systems won't. While Grok models prior to Grok 3 hedged on political subjects and won't cross certain boundaries, it's unclear whether xAI has achieved its goal of shifting Grok closer to politically neutral. One study found that Grok leaned to the political left on topics like transgender rights, diversity programs, and inequality.

The release of Grok 3 marks a significant development in the AI industry, with implications for the future of AI research, development, and application. As xAI continues to push the boundaries of AI capabilities, it remains to be seen how Grok 3 will be received by users and the broader tech community.