Grok keeps updating after xAI disbandment as Musk announces a new model

xAI解散，但Grok还在上新，马斯克官宣新模型

Elon Musk said the 1.5T-parameter Grok V9-Medium has finished training, will enter reinforcement learning in a few days, and is planned for release in two to three weeks. Grok Build supports up to 8 parallel sub-agents, a 256K-token context window, Plan Mode, Arena Mode, MCP, and ACP.

Why it matters: HKR-H/K/R all pass, but this is a Grok V9-Medium preview before RL and release, with no benchmarked capability yet. That fits a strong model-race/product update at 82, featured but not p1.