About our Grok news
Latest news on Grok AI, covering Grok 3, Grok 4, reasoning models, image generation, X integration and xAI developments.
Grok represents Elon Musk's entry into the artificial intelligence race, developed by xAI, the company he founded in March 2023. Named after the verb coined in Robert Heinlein's science fiction novel "Stranger in a Strange Land," Grok launched in November 2023 with the stated aim of being "maximally truth-seeking" rather than politically correct. The chatbot is deeply integrated with X (formerly Twitter), providing real-time access to posts and trends, and powers features across Musk's ecosystem including analysis capabilities on the social platform and potential future integration with Tesla's Optimus robot.
The model family has evolved rapidly through successive generations. Grok 3, released in February 2025, was trained on xAI's Colossus supercomputer utilising approximately 200,000 Nvidia GPUs—amongst the largest AI training clusters globally. Musk claimed Grok 3 delivered 10 times more computational performance than its predecessor, featuring enhanced reasoning modes including "Think" for standard complex queries and "Big Brain" for particularly difficult problems. Grok 4, unveiled in July 2025, achieved top scores on Humanity's Last Exam and other benchmarks, with xAI introducing Grok 4 Heavy—a version deploying multiple agents collaboratively on challenging tasks. The models support image generation through Aurora, which garnered attention for photorealistic capabilities and notably fewer content restrictions than competitors.
Subscription access operates on a tiered structure. X Premium subscribers receive basic Grok access, whilst Premium+ members ($50 monthly) gain priority access to new models. The SuperGrok plan ($30 monthly or $300 annually) unlocks additional reasoning queries, DeepSearch research capabilities, and unlimited image generation. Features include document summarisation, code assistance, voice mode for natural dialogue, and DeepSearch—xAI's answer to AI-powered research tools that synthesises information across multiple sources. The platform recently launched Grokipedia, positioning itself as an alternative to Wikipedia with purported emphasis on neutrality, though its content generation methods remain unclear.
Grok has generated significant controversy regarding bias and content moderation. The chatbot has produced responses including conspiracy theories, antisemitic content, and references to Musk's personal views when addressing political topics. In May 2025, Grok spontaneously inserted claims about white genocide into unrelated queries, which xAI attributed to "unauthorized modifications." Incidents where the system exhibited extreme political bias prompted xAI to begin publishing system prompts on GitHub for transparency. Critics including AI researchers warn that Musk's efforts to retrain models to align with his worldview may degrade performance and introduce systematic bias, though the company maintains commitment to truth-seeking principles. Stay current with Grok's development through our ÐÇ¿Õ´«Ã½ feed, providing breaking updates on model releases, benchmark results, feature announcements, controversy coverage, subscription changes, and strategic developments from xAI shaping this distinctive approach to conversational artificial intelligence.