Ultra-fast inference
Groq runs popular open LLMs on its own LPU hardware for record-breaking inference speeds.