[Musk: Grok V9-Medium Model with 1.5 Trillion Parameters Expected to Launch in Two to Three Weeks] Musk posted that xAI has completed training the Grok foundation model V9-Medium (1.5T parameters), which is currently undergoing fine-tuning, with reinforcement learning set to begin in a few days. It is expected to be made publicly available in 2-3 weeks. Compared to the current 0.5T V8-small model, which handles all Grok online traffic, the new model will bring 'significant improvements' in complex programming tasks. He also mentioned that the existing 0.5T model is planned to be open-sourced by the end of this year and is expected to remain practical. This round of training has incorporated a large amount of Cursor-related data, with more to be added in the future.