SpaceXAI’s announcement on X introduces Grok 4.5 as the company’s first model “trained specifically for coding and agents,” with the post claiming it was trained with Cursor and built for “frontier intelligence” at “leading speeds and cost efficiency.”
The company also states that Grok 4.5 is aimed at “real-world engineering,” including large codebases and long-running tasks that span multiple repositories, hundreds of skills, and different tools. SpaceXAI further claims the model is more efficient than comparable systems, saying it uses half as many tokens per task, delivers higher tokens per second, and costs less than half as much, at $2 per million input tokens and $6 per million output tokens.
SpaceXAI says Grok 4.5 is available today in the SpaceXAI console, Grok Build, and Cursor, with all limits reset so users can start for free. The company added that the model is not yet available in the EU, with availability expected later this month.
The announcement also included a benchmark chart labeled “DeepSWE 1.0,” where Grok 4.5 registered 62% pass@1 within each model provider’s harness. That placed it behind Fable max at 66.1% and GPT 5.5 xhigh at 64.3%, while ahead of Opus 4.8 max at 55.8% and Opus 4.7 max at 54%.
Early replies to the post focused on access and rollout, with users asking about Cursor support, CLI availability, and the EU restriction.
Source: SpaceXAI on X
