xAI rebrands to SpaceXAI and ships Grok 4.5 frontier coding model

xAI, rebranded SpaceXAI after SpaceX's acquisition, shipped Grok 4.5 on July 8, an 'Opus-class' model built for coding, agentic tasks and knowledge work. Elon Musk amplified the launch on July 13, claiming Grok is 'moving faster than any competitor right now.' The model is priced aggressively at $2 per 1M input tokens — less than half of Anthropic's comparable Opus — and is served free through Cursor, positioning it squarely against Claude Sonnet 5 and GPT-5.6 for engineering work.

The distinguishing feature is deep Cursor data integration: Grok 4.5 was trained alongside the popular AI code editor, which xAI argues gives it an edge on real-world coding workflows. It is available through the xAI API, Grok Build and the xAI console, and claims competitive benchmark performance with greater speed and token efficiency than rivals.

Validation came from Perplexity CEO Aravind Srinivas, who said Perplexity integrated Grok 4.5 inside Perplexity Computer within hours because 'it scored the best on our evals and was the most cost effective option,' with zero-data-retention available from day one — a signal enterprise buyers care about.

Competitively, the pricing and Cursor distribution make Grok 4.5 a genuine threat in the developer segment, the same week Anthropic and OpenAI both refreshed their flagships. Skeptics note Musk's 'fastest' claims are marketing and that independent long-horizon coding benchmarks are still thin. Watch whether the SpaceXAI rebrand brings compute and capital advantages, and whether Cursor's tight coupling raises lock-in concerns among developers.