Introducing Grok 4.5: SpaceXAI's Smartest Model for Coding and Agentic Tasks

Grok 4.5

We are thrilled to launch Grok 4.5, our most intelligent model designed to excel at coding, complex engineering, and office work. Trained on massive datasets and optimized for efficiency, it outperforms leading models on benchmarks like DeepSWE and Harvey's Legal Agent. With speeds of 80 TPS and significantly lower token costs, Grok 4.5 delivers high-quality results for building apps, analyzing data, and creating documents faster than ever before.

"Grok 4.5 is incredibly capable at coding, from challenging Rust and C/C++ tasks to end-to-end app building from prompt to production."