Summary
SpaceXAI (formerly xAI) launched Grok 4.5, built on a ~1.5-trillion-parameter V9 foundation and supplemented with Cursor coding data, positioning it for coding, agentic tasks, and knowledge work. It is available in Grok Build, in Cursor on all plans, and through the SpaceXAI console at $2 per million input and $6 per million output tokens, though not yet in the EU.
What changed
SpaceXAI shipped Grok 4.5 across Grok Build, Cursor, and its console, priced at $2/$6 per million tokens, and reported it resolves SWE-Bench Pro tasks at roughly 4x lower output-token cost than a top competing model.
Why it matters
A frontier coding model shipping directly inside Cursor on all plans — and trained partly on Cursor data — tightens the loop between model labs and the editors that distribute them. Aggressive token-efficiency and pricing claims intensify the cost competition among agentic coding models where per-task token spend increasingly decides tool economics.
Evidence excerpt
Grok 4.5 is available in Grok Build, in Cursor on all plans, and through the SpaceXAI console, priced at $2 per million input tokens and $6 per million output tokens.