SpaceXAI launches Grok 4.5 AI model for coding, agentic tasks

SpaceXAI announced on Wednesday the launch of its Grok 4.5 AI model, which the company hails as its most intelligent release to date, specifically designed for coding and agentic tasks.

The Grok 4.5 model underwent training using tens of thousands of Nvidia GB300 graphics processing units. SpaceXAI emphasized meticulous data filtering, deduplication, and quality scoring during the training process. Cursor, a popular AI coding agent, confirmed its partnership with SpaceXAI for training Grok 4.5. This move follows SpaceX's announcement last month of its intent to acquire Anysphere, the startup behind Cursor, in a $60 billion all-stock deal aimed at expanding its enterprise AI tools market presence.

Grok 4.5 is immediately accessible via SpaceXAI's AI coding agent, Grok Build, within Cursor, and through the SpaceXAI console, the company's developer portal, using an API key. Availability in the European Union is anticipated by mid-July.

“

We've partnered with SpaceXAI to train Grok 4.5.

The pricing for Grok 4.5 is set at $2 per million input tokens and $6 per million output tokens. SpaceX CEO Elon Musk described the model as "an Opus-class model, but faster, more token-efficient and lower cost." This positions Grok 4.5 as a competitive offering against rivals such as Anthropic's Claude Opus 4.8, which is priced at $5 per million input tokens and $25 per million output tokens, and OpenAI's GPT-5.6 Luna, priced at $1 per million input tokens and $6 per million output tokens.

Input tokens refer to the data sent to an AI model, while output tokens are the responses generated by the model. OpenAI is scheduled to publicly launch its most advanced AI model, GPT-5.6, on Thursday, following a delay attributed to U.S. government requests concerning national security implications of powerful AI technologies.

“

It is an Opus-class model, but faster, more token-efficient and lower cost.