RTX 4060 Ti 16GB llama.cpp: gpt-oss 20B MXFP4 decode at 128k ctx (Hardware Corner)
31.1 tok/s token generation, batch 1, per-request, ctx 128k
Fields
id
benchmarks/hc-4060ti-gptoss-20b-mxfp4-128k-tgMetric
decode_tok_s
Value
31.1
Unit
tok/s token generation, batch 1, per-request, ctx 128k
Model
gpt-oss 20B
Methodology
Hardware Corner per-GPU context-tier table (March 2026 update), single RTX 4060 Ti 16GB, llama.cpp GGUF, batch-1 per-request. gpt-oss 20B MXFP4, context depth 128k. Prefill column is prompt-processing tok/s; generation column is decode tok/s. NOT an aggregate across concurrent requests. The page does not reprint the llama-bench build hash (the site's sibling ranking page pins CUDA 12.8 / Q4_K_XL).
Measured by
third_party
Reproducible
yes
notes
Paired prefill is 780.2 tok/s in the same table cell. Per-request, not aggregate.
Notes
Paired prefill is 780.2 tok/s in the same table cell. Per-request, not aggregate.