SpaceXSI releases Grok 4.7 in Kiro

SpaceXSI has released Grok 4.7 in Kiro, offering improved performance in coding, long-horizon tasks, and knowledge work. According to SpaceXSI and independent testing by Artificial Analysis, the new model features higher benchmark scores, reduced hallucination rates, and enhanced self-checking capabilities across a 500K token context window. It is available across multiple Kiro interfaces with four reasoning effort levels and a new safeguard stack. Open article →