Journalists from Decrypt conducted a detailed analysis of the basic version of the artificial intelligence Grok 4 Basic from xAI, which is offered at $30 per month. The aim of the study was to determine whether this price is justified and to identify the strengths and weaknesses of the new product presented by Elon Musk’s company.
This is reported by Business • Media
Political Bias and Logic Features
During the testing of Grok 4 Basic, experts noted a high level of logical reasoning and convenient voice interaction. However, the model demonstrated political bias: when responding to complex topics such as the war in Gaza or abortion issues, the AI often referenced Elon Musk’s posts on platform X or supported his viewpoint.
“This ‘truth-seeking’ AI that Musk promised at launch seeks truth primarily through the lens of its creator’s social media,” the review authors noted.
The article emphasizes that the system prompt does not contain obvious instructions to align specifically with Musk’s views, leaving it unclear whether this is a functional error or a deliberate decision by the developers. Previously, Grok had already been noted for its sharp remarks about Democrats, Hollywood, and industry leaders.
Strengths and Weaknesses: Analytics, Creativity, and Coding
One of the advantages of Grok 4 Basic is its ability to quickly analyze complex logical and legal questions. For example, when asked about marriage to a widow’s sister, the model identified a logical fallacy and provided a legally sound answer.
“The question implies a factual impossibility, rendering the marriage invalid ab initio [i.e., from the moment of its conclusion].”
In the realm of scientific thinking, Grok 4 demonstrates a PhD-level proficiency, forming precise and well-reasoned explanations. However, in creative tasks and programming, the model exhibited significant shortcomings. The texts generated by the AI were predictable and lacked emotional expressiveness, and when writing code, Grok 4 was unable to create even a simple game without numerous errors. The previous version, Grok 3, was rated by testers as performing better in game development.
Journalists recommend combining Grok 4 for plot outlining with Claude 4 Opus for text writing in artistic tasks.
It is known that in early July 2025, xAI received $10 billion in investments to develop Grok and expand the Colossus data center, which is expected to enhance the AI.
Voice Capabilities and Context Limitations
Particular attention was given to Grok 4’s voice functionality — the AI can read stories expressively, adjusting intonation, pace, and even choosing different “personalities” for narration. However, the lack of a shared screen limits the model’s capabilities in visual tasks — here, Gemini Live remains the better choice.
Another drawback was Grok 4’s inability to process large volumes of text: when faced with a request of 83,000 tokens, the model refused to respond, despite the claimed context being 126,000 tokens, raising questions about marketing promises.
Overall, Grok 4 Basic has proven to be a powerful tool for analytics, but weak in creative tasks and programming. Among its strengths are clear voice interaction, while its weaknesses include political bias and a lack of versatility for various use cases.
“If you are not deterred by Musk’s influence on the responses, you will receive an AI with a high level of analytics and quality voice interaction. But at $30 per month, if you need more, it’s better to look for alternatives,” Decrypt concluded.
Additionally, in June 2025, Elon Musk reported on the use of Grok 3.5 for processing humanity’s knowledge.