1. Reasoning
StrongAccuracy and coherence when solving multi-step problems and explaining conclusions.
Grok 4 and later named models show strong capability on complex reasoning and extended knowledge-work tasks, with reported strengths in multi-step problem solving and technical work. Performance is not uniform across task types, and evidence about newer models does not establish that every version or app session behaves alike.
Evidence confidence: medium