NVIDIA Groq 3 LPX hits 3,431 TPS on Vera Rubin, enhancing AI efficiency with high-interactivity tasks.
NVIDIA's latest product, the Groq 3 LPX, has raised the bar for AI inference, achieving an impressive 3,431 tokens per second on a benchmark with 100,000 contexts. This performance milestone sets a new standard for high-interactivity workloads in the field.