On September 2, Elon Musk announced that the next-generation large language model "Grok 4.7," developed by his subsidiary SpaceXAI, will be officially released on September 12, 2026.
Model Releasesイーロン・マスクSpaceX AIGrok 4.7
SpaceXAI to release next-generation model "Grok 4.7" on September 12, expanding to 2.1 trillion parameters
This article is a translation. Read the Japanese original
The new model will be built on a foundation architecture of 2.1 trillion parameters. This represents a 40% increase in scale compared to the previous generation, Grok 4.6 (1.5 trillion parameters). Musk stated that Grok 4.7 will outperform Grok 4.6 in every aspect and will have higher token efficiency.
While the inference speed is expected to decrease slightly due to the increase in parameter scale, the computational efficiency per token is said to improve. Furthermore, he mentioned that the model is in a supplemental training phase integrating SpaceX's corporate data, aiming to build a model with an advantage in real-world engineering capabilities.
Sources
- マスク氏、Grok 4.7を10日後に発表へ パラメータ数2.1兆で「全モデル超え」と豪語 - BigGo ファイナンス （Google News: Grok、2026-09-02）