T4 GPU cloud prices

NVIDIA T4 is commonly evaluated for LLM inference. The Turing T4 (16 GB, 70 W) — a low-power inference staple available almost everywhere.