NVIDIA DGX Station serves Qwen3.8-27B at 2,713+ tok/s peak
NVIDIAAI · x · 2026-08-20
Benchmarks show the NVIDIA DGX Station serving the Qwen3.8-27B model at BF16 full weight achieves over 2,713 tok/s peak aggregate throughput and 88 tok/s single-stream, effectively bringing datacenter-level performance to a home setting.
More from Infra
- Haight Street Compute Market: a one-day trading floor for GPU deals in SF — ctjlewis · 2026-10-16
- Temporal Serverless Workers on Cloud Run: queue-driven autoscaling, scale to zero — rseroter · 2026-10-16
- LMA brings day-0 model support to TPU for the first time, spanning NVIDIA, AMD, Intel via SGLang — ying11231 · 2026-10-16
- Semtech's analyst-day slide cleanly explains the four signal-integrity impairments — BenBajarin · 2026-10-16
- Fortytwo Engine boosts Qwen3.8-27B accuracy 7% and cuts tokens 25% via Swarm Inference — rohanpaul_ai · 2026-10-16
- Berkeley scholars: concentrated chip supply chain makes a hardwired frontier AI pause feasible — wfithian · 2026-10-16