SpaceXAI launches Grok 4.5, built for coding and agent tasks

SpaceXAI has released Grok 4.5, its first model designed specifically for coding and agent tasks. Built with Cursor, Musk calls it an 'Opus-level model'.

Grok 4.5 handles large codebases and long tasks across multiple repositories. Trained on tens of thousands of NVIDIA GB300 GPUs, it uses reinforcement learning on hundreds of thousands of tasks focused on multi-step software engineering. It uses half the tokens per task as leading models, outputs 80 tokens per second, and costs $2 per million input tokens and $6 per million output tokens, half the cost of similar models.

Beyond coding, it handles Office tasks like building complex Excel models, designing PowerPoint slides, and writing Word documents. Available now in SpaceXAI console, Grok Build, and Cursor with free usage, but not yet in the EU—expected mid-July.