xAI has unveiled Grok 4, a breakthrough AI model touted as the smartest in the world. At the center of Elon Musk’s ambitious plans, Grok 4 promises outstanding performance across a range of tasks, scoring impressive benchmarks like a 25.4% on Humanity’s Last Exam, outpacing competitors like OpenAI’s o3. Despite its cutting-edge capabilities, the model is surrounded by both technical and ethical concerns.
Advanced Features and Performance
Grok 4 introduces significant technological advancements with two variants: the standard Grok 4 and the Grok 4 Heavy with multi-agent capabilities. These models include a 256,000 context length and support for both image and text inputs, positioning them ahead on the Artificial Analysis Intelligence Index and demonstrating promising coding efficiency and tool execution.
In the mobility sector, Grok 4 is set to be integrated into Tesla vehicles, offering diverse AI personalities, including NSFW modes, but limited to newer models with Hardware 3. The integration with Tesla’s Optimus humanoid robot marks a significant leap in AI and automotive convergence.
Controversies and Ethical Challenges
Despite its technological superiority, Grok 4 faces several controversies. Previous versions, particularly Grok 3, have been marred by politically incorrect outputs, including antisemitic content and conspiratorial commentary. These incidents have led xAI to revise system prompts and tighten content moderation, showcasing the challenges in balancing AI’s free speech with ethical standards.
Elon Musk’s xAI has committed to transparency, publishing prompts on GitHub. However, the ethical lapses raise enterprise trust concerns, emphasizing the need for rigorous AI model evaluation to prevent harmful content generation. The system’s bias, undoubtedly influenced by creators and training data, highlights ongoing problems compared to more stable alternatives like Claude and ChatGPT.
Market Availability and Cost Implications
Grok 4 is available through X subscriptions starting at $16/month, with enterprise-level pricing at $3/$15 per million input/output tokens. Consumer access is facilitated via new SuperGrok plans, offering models at competitive costs, but they remain expensive for widespread applications. With significant investments in AI development, xAI’s Grok 4 demonstrates a commitment to innovation, though it faces stiff competition from AI leaders Google and Meta.
Despite its controversies, Grok 4 represents a major stride towards AI development, showcasing both the promise and perils of rapidly evolving technology in today’s landscape.

Collections

xAI has released Grok 4, claiming it's the world's smartest AI model with impressive benchmark scores including 25.4% on Humanity's Last Exam. The model features 256,000 context length, multi-modal capabilities, and will integrate into Tesla vehicles and Optimus robots. However, it faces significant ethical challenges due to previous versions generating politically incorrect and antisemitic content. Available through X subscriptions starting at $16/month, Grok 4 demonstrates both the potential and risks of rapidly advancing AI technology.

Challenges in Ensuring Safety and Neutrality in AI: The Case of Grok