SpaceXAI is releasing Grok 4.5, its most advanced AI model to date, following its public debut and the acquisition of AI coding firm Cursor. This new model is primarily designed as a tool for coding and agentic tasks, rather than a consumer chatbot. Elon Musk describes Grok 4.5 as "maximally truth-seeking" compared to its competitors. The company claims it surpasses comparable models in engineering and knowledge work, being faster, more token-efficient, and less expensive than Anthropic's Opus model. Grok 4.5 is now accessible through Grok Build, Cursor, and the SpaceXAI console, excluding the EU. While it outperforms some OpenAI and Anthropic models in speed, price, and performance, it has not yet matched the very latest from those companies. SpaceXAI expects to surpass these leading models in the near future. The model is competitively priced, significantly cheaper than Anthropic's Opus 4.8 and OpenAI's GPT 5.6 Luna on a per-token basis. Intriguingly, Grok 4.5 was trained using compute capacity that SpaceXAI also leases to competitors like Anthropic and Google. This situation presents a potential conflict as SpaceXAI's increasing compute demands might necessitate a choice between internal use and external leasing for revenue. OpenAI is also releasing its GPT 5.6 model and new voice models concurrently. Ultimately, SpaceXAI is positioning Grok 4.5 as a cost-effective, business-oriented AI solution that, despite currently trailing the top-tier models of rivals, aims to be a strong contender in the market.