The newest iteration of the Grok AI assistant from SpaceXAI provides double the speed of competing systems and enhanced ability to manage longer context windows. SpaceXAI set the pricing for its 4.7 model at $2 per 1 million input tokens and $6 per 1 million output tokens. This pricing structure reduces the cost of using comparable high-end models by 50% and improves the system's capacity for self-verifying its own work.

The debut occurred via an official blog post on Sept. 21, following a brief leak on the OpenCode Zen platform. The launch comes during a period of intense competition in the frontier model sector, with industry observers anticipating updated releases from OpenAI and Anthropic later this week.