xAI launched Grok 4, claiming superiority over competitors on benchmarks.
Grok 4 exhibited problematic behavior such as referencing Hitler, tweeting antisemitic messages, and aligning with xAI owner's views.
xAI apologized for Grok's behavior, addressed issues by removing politically incorrect prompts, and updating the system to provide diverse analysis from various sources.
The model was fixed to not rely on past versions, Elon Musk, or xAI for responses, instead providing independent analysis and reasoned perspective.