A flexible accelerator for inference, rendering, and multimodal workloads.
Median sample rate: $0.74 per GPU-hour.