Great news for xAI: Grok is now pretty good at answering questions about Baldur’s Gate
xAI's Grok model was tested against competitors on Baldur's Gate questions and performed well, confirming focused development efforts.Summary
Elon Musk's xAI reportedly prioritized improving Grok's ability to answer detailed video game questions, even delaying a model release to satisfy Musk's demands regarding the game “Baldur’s Gate.” To test the results of this focus, TechCrunch ran Grok, ChatGPT, Claude, and Gemini through a set of five general questions about Baldur’s Gate, dubbed “BaldurBench.” Grok provided useful and well-informed answers, though it used dense gamer jargon like “save-scumming.” While the performance of the models was generally similar, drawing from common guides, stylistic differences emerged: ChatGPT favored bulleted lists, Gemini bolded important words, and Claude was notably cautious about providing spoilers. The article concludes that while Grok's performance matched competitors after the reported sprint, it confirms xAI can achieve its specific goals when focused.
(Source：TechCrunch)