Grok is a conversational chatbot powered by xAI's large language model (LLM) called Grok-1 and generates automated responses based on data from the web, real-time news from X, and feedback from human assistants.

Upon its original launch last year, Grok was only accessible to X's Premium subscribers, which make up fewer than 1% of the platform's total user base. This includes Grok-2 and Grok-2 mini, two newer LLMs that feature an in-app image generator and upgraded performance.

Eventually, the social media company wanted to bring more users to Grok, testing a "freemium" version of the chatbot, which has now rolled out fully, offering users the ability to ask Grok questions, generate images via an enhanced image-generator called Aurora, analyze posts, and generate humorous responses through an "Unhinged Fun" mode.

Making Grok more widely available may increase X's chances of competing closely with OpenAI's ChatGPT, Google Gemini, Microsoft Copilot, and Anthropic's Claude. Like these leading generative-AI chatbots, Grok is working on launching a standalone app, according to a report from The Wall Street Journal./p> <p]>What currently sets Grok apart from other chatbots is a lack of restrictions and parameters regarding what users can automate, including impressive images of public and copyrighted figures carrying out unflattering, and oftentimes offensive, acts.

Upon the time of this article, Aurora has been removed after briefly opening upon Grok's free network-wide launch, potentially for an accidental leak, or TechCrunch's ability to generate an image of "a blooded [Donald] Trump," which other image generators are restricted from creating.

In addition, it is not clear how Aurora was trained.

X has yet to reveal whether Musk's AI company, xAI, which recently received $6 billion in funding, trained Aurora itself, or whether it was created in collaboration with a third party.