board members over his divided attention due to the Tesla CEO’s political activities. Musk claimed Grok 4 achieves “PhD levels in every subject” with “no exceptions” when handling academic questions. The model demonstrated superior performance on ARC-AGI benchmarks, scoring 66.6% on version 1 and 15.9% on the more challenging version 2 tests. These results significantly exceed competing models, with OpenAI’s o3 achieving 60.8% on ARC-AGI v1 and 6.5% on v2. Anthropic’s Claude Opus 4 reached 35.7% and 8.6% respectively, while Google’s Gemini 2.5 Pro scored 41.0% and 4.9%. The benchmark measures artificial general intelligence capabilities through reasoning tasks, with Grok 4’s performance suggesting substantial advancement in AI reasoning abilities. However, the model operates at higher computational costs per task compared to competitors. Wedbush Securities analyst Dan Ives recently called for the company’s board to facilitate a potential merger between Tesla and xAI, arguing it could create an “AI stalwart” supporting Tesla’s robotaxi and humanoid robotics initiatives. Musk previously stated that autonomous and robotics will “overwhelmingly dominate” Tesla’s future financial success.