Elon Musk’s xAI has announced that it will release two new versions of its AI model, Grok 4.6 and Grok 4.7, in August, continuing the company’s rapid pace of AI development.
The first release, Grok 4.6, is scheduled for August 7. The model will feature 1.5 trillion parameters and improved Supervised Fine-Tuning (SFT) and Reinforcement Learning (RL) capabilities, enabling it to better understand user prompts and handle complex tasks more effectively.
A few weeks later, xAI will introduce Grok 4.7, powered by 2.1 trillion parameters. According to Elon Musk, the new model will outperform Grok 4.6 across nearly every benchmark. While it may be slightly slower, it is expected to use tokens more efficiently and generate higher-quality responses.
Earlier this month, xAI unveiled Grok 4.5, a version designed primarily for developers and engineers. Its key focus areas include coding, AI agent orchestration, and solving advanced reasoning tasks.
With back-to-back model releases, xAI continues to accelerate the development of the Grok family, strengthening its competition with leading AI companies such as OpenAI, Anthropic, and Google.
