SpaceXAI launched Grok 4.5 on July 8, positioning it as the company’s strongest model for coding, agentic tasks, and knowledge work.
The launch page says Grok 4.5 was trained with Cursor and is available in Grok Build, Cursor on all plans, and the SpaceXAI API console. It is not yet available in the EU across SpaceXAI products or the API console, with EU availability expected in mid-July.
SpaceXAI prices Grok 4.5 at $2 per million input tokens and $6 per million output tokens. The company also says the model is served at 80 tokens per second and uses fewer output tokens than comparable leading models on SWE Bench Pro tasks.
The pitch is speed and cost, not only rank
SpaceXAI’s launch page includes benchmark claims across DeepSWE, SWE Marathon, Terminal Bench 2.1, and SWE Bench Pro. It says Grok 4.5 reaches 62.0% on DeepSWE 1.0, 83.3% on Terminal Bench 2.1, and 64.7% on SWE Bench Pro.
Those figures should be treated as vendor-reported launch evidence, not independent ranking proof. They still show how SpaceXAI wants Grok 4.5 judged: less as a general chatbot upgrade and more as an engineering model competing on task completion, latency, and token efficiency.
That framing fits the market. Coding agents are expensive when they run long trajectories, call tools repeatedly, and produce large patches. A model that can solve a similar class of tasks with fewer tokens can be more valuable than a model with a small benchmark edge and much higher run cost.
Availability is already a product strategy
Grok 4.5 is not launching as a model card alone. SpaceXAI is putting it into Grok Build, Cursor, and the API console immediately.
That means the first adoption path is workflow distribution. Developers can try it inside a coding environment, product builders can use Grok Build, and API users can wire it into their own tools. The EU exclusion is the practical caveat for global teams.
The next thing to watch is independent evaluation. If Artificial Analysis, coding-agent vendors, and enterprise pilots find that the token-efficiency claims hold under real repositories, Grok 4.5 becomes a cost pressure point for other frontier coding models.
