ai-tools model-release coding-with-ai developer-tools industry

SpaceXAI and Cursor Launch Grok 4.5 — A Coding-First AI Model

SpaceXAI and Cursor publicly launch Grok 4.5, a joint AI model built for coding, legal, and finance tasks. Available today.

AI Tutorials · 8 July 2026 · 2 min read

Quick answer

SpaceXAI and Cursor have publicly launched Grok 4.5, the first AI model built jointly since SpaceX acquired Cursor for $60 billion. It's a mixture-of-experts model trained on trillions of tokens of real coding interactions, designed for software engineering, legal, and financial tasks. It's available now in Cursor on all plans and through the SpaceXAI console.

product-launch

SpaceXAI and Cursor Launch Grok 4.5 — A Coding-First AI Model

SpaceXAI and Cursor have publicly launched Grok 4.5, the first AI model built jointly since SpaceX acquired Cursor in a deal valuing the coding startup at $60 billion. After weeks in private beta at SpaceX and Tesla, the model is now available to everyone.

What Makes Grok 4.5 Different

This isn’t a general-purpose chatbot. Grok 4.5 is a mixture-of-experts model trained on trillions of tokens of real Cursor user interactions — actual coding sessions, debugging workflows, and software development tasks. Cursor CEO Michael Truell says the model was trained from scratch on Colossus, xAI’s supercomputer.

The result is a model built specifically for working across large codebases. It can generate production-ready code, debug software, plan development tasks, and handle multi-step coding workflows. SpaceXAI also expanded its capabilities into legal analysis and financial document processing, positioning it as a professional tool rather than a consumer chatbot.

SpaceXAI claims roughly 2x the token efficiency of comparable frontier models, meaning it can solve tasks in under half the number of steps. Elon Musk calls it “Opus-class but cheaper.”

Pricing and Availability

Grok 4.5 is available today in three places: directly within Cursor on all subscription plans, in Grok Build (SpaceXAI’s developer platform), and through the SpaceXAI API console.

Pricing undercuts the competition at $2 per million input tokens and $6 per million output tokens. A faster premium tier runs $4/$18. For comparison, Claude Opus and GPT-5 class models typically cost significantly more per token.

The model is not yet available in the EU due to regulatory considerations.

What This Means for You

If you use Cursor for coding, Grok 4.5 is now another model option alongside Claude and GPT. The tight integration with Cursor’s editor — combined with training data from millions of real coding sessions — could give it an edge for day-to-day development work.

For developers choosing between AI coding tools, this launch signals that the coding AI race is accelerating fast. The SpaceX-Cursor combination now has the compute power, the user data, and a competitive model to challenge Anthropic and OpenAI directly.

Frequently asked questions

What is Grok 4.5?

Grok 4.5 is a new AI model built by SpaceXAI and Cursor together. It's a mixture-of-experts model trained on real developer coding sessions, designed primarily for software engineering, legal analysis, and financial tasks rather than general chatbot use.

How much does Grok 4.5 cost?

Grok 4.5 is priced at $2 per million input tokens and $6 per million output tokens. A faster premium version costs $4 per million input tokens and $18 per million output tokens. It's also available directly within Cursor on all subscription plans.

How does Grok 4.5 compare to Claude and GPT?

SpaceXAI claims Grok 4.5 is 'Opus-class' in coding performance but cheaper, with roughly 2x the token efficiency of comparable models. Independent benchmarks are still emerging, but the model is trained specifically on real coding workflows from Cursor's user base.

Is Grok 4.5 available in my country?

Grok 4.5 is available globally through the SpaceXAI console and Cursor, but it is not yet available in the European Union due to regulatory considerations.