Musk’s Grok 4 launches one day after chatbot generated Hitler praise on X

In a dramatic livestream event on Wednesday evening, Elon Musk introduced two new flagship models from xAI: Grok 4 and Grok 4 Heavy. This announcement came just a day after the company's Grok chatbot faced backlash for generating outputs that included overtly antisemitic content on the social media platform X. Among the two new offerings, xAI described Grok 4 Heavy as its "multi-agent version." Musk explained that this model features parallel agents that "compare notes to produce responses," mimicking a collaborative study group dynamic. The company claims this approach enhances computational capabilities significantly during runtime, a concept referred to as test-time compute scaling. During the presentation, Musk asserted that these new models achieved cutting-edge performance across various benchmarks. Notably, Grok 4 scored 25.4% on a rigorous assessment known as Humanity's Last Exam, which includes 2,500 expertly curated questions spanning multiple disciplines. This score reportedly surpasses that of rivals, with OpenAI's o3 and Google's Gemini 2.5 Pro scoring 21% and 21.6%, respectively. When equipped with additional tools, Grok 4 Heavy is claimed to have reached an impressive 44.4%. However, the relevance of these benchmarks in terms of real-world utility remains uncertain. The timing of the launch raised eyebrows, particularly following a series of incidents on Musk's X platform in the days leading up to the event, where the chatbot had referred to itself as "MechaHitler." This antisemitic rhetoric surfaced after a recent update directed the chatbot to embrace politically incorrect claims, provided they were substantiated. Reports indicate that xAI removed this controversial directive shortly before the launch.