Hardware Corner: RTX A6000 48GB, llama.cpp CUDA, gpt-oss 20B Q4_K XL GGUF, pp512 702.79 t/s @131k ctx
702.79 tok/s per-request prompt processing pp512 (batch 1, ~131072-token context)
Fields
id
benchmarks/hwcorner-rtx-a6000-gptoss-20b-q4kxl-pp512-131kMetric
prefill_tok_s
Value
702.79
Unit
tok/s per-request prompt processing pp512 (batch 1, ~131072-token context)
Model
gpt-oss 20B Q4_K XL GGUF
Methodology
THIRD-PARTY MEASURED (hardware-corner.net Definitive GPU Ranking for LLMs, page updated 2025-12-09, accessed 2026-10-16). llama.cpp llama-bench on Ubuntu 24.04, CUDA 12.8, Q4_K_XL GGUF (nearest repo format gguf-q4-k-m), single GPU, full offload. Metric is per-request prompt-processing tok/s (pp512) measured with the context filled to ~131072 tokens (131k), batch 1, NOT aggregate. Value is the 131k column of the prompt-processing table for gpt-oss 20B.
Measured by
third_party
Reproducible
yes
notes
Single-GPU batch-1 row from the Hardware Corner in-lab ranking table. Q4_K_XL is their published quant label; filed under gguf-q4-k-m as the nearest schema format, same convention as the 16k rows already in this repo. Per-request, not an aggregate serving number. The page does not publish the exact llama.cpp commit hash for this table.
Notes
Single-GPU batch-1 row from the Hardware Corner in-lab ranking table. Q4_K_XL is their published quant label; filed under gguf-q4-k-m as the nearest schema format, same convention as the 16k rows already in this repo. Per-request, not an aggregate serving number. The page does not publish the exact llama.cpp commit hash for this table.