Hardware Corner Qwen3.8-27B Q4_K_S: dual RTX 5060 Ti 16GB (2x15849 MiB) tg 22.47 t/s @4k ctx
22.47 tok/s per-request decode (batch 1, 4k context)
Fields
id
benchmarks/hc38-dual-rtx-5060-ti-qwen38-27b-tg-4kMetric
decode_tok_s
Value
22.47
Unit
tok/s per-request decode (batch 1, 4k context)
Model
Qwen3.8 27B Q4_K Small GGUF
Methodology
THIRD-PARTY MEASURED (hardware-corner.net 'We Tested Qwen3.8 27B', accessed 2026-10-15). llama.cpp build 153d324bc (10364), Ubuntu 24.04, NVIDIA driver 595.58.03, CUDA 12.8, Qwen3.8-27B Q4_K Small GGUF (16.68 GiB, 27.32B params; llama.cpp reports arch 'qwen35'), Flash Attention on, MTP off, full GPU offload. dual RTX 5060 Ti 16GB (2x15849 MiB). pp = prompt-processing tok/s, tg = token-generation tok/s at stated prefilled context. Per-request batch-1.
Measured by
third_party
Reproducible
yes
notes
TWO-card run, tensor-split across 2x 5060 Ti; NOT single-card. Marked in methodology.
Notes
TWO-card run, tensor-split across 2x 5060 Ti; NOT single-card. Marked in methodology.