Speed sweep

qwen3-6-35b-a3b-nsc-ace-saber-rocmfp4-ryzen-ai-max-plus-395-128gb-llama-cpp-tp1-sweep

qwen3-6-35b-a3b-nsc-ace-saber-rocmfp4-ryzen-ai-max-plus-395-128gb-llama-cpp-tp1-sweep

Record

Recipe
qwen3-6-35b-a3b-nsc-ace-saber-rocmfp4-ryzen-ai-max-plus-395-128gb-llama-cpp-tp1
Measured
2026-06-07T19:13:33.476Z
Points
1

Measured speed

ConcurrencyContextPrefillDecodeTTFT msStatus
1500,000772137.9observed

Remaining fields

Identity, launch, related records, and measured speed are shown above. This is the rest of the normalized record.

id
qwen3-6-35b-a3b-nsc-ace-saber-rocmfp4-ryzen-ai-max-plus-395-128gb-llama-cpp-tp1-sweep
measured at
2026-06-07T19:13:33.476Z
recipe id
qwen3-6-35b-a3b-nsc-ace-saber-rocmfp4-ryzen-ai-max-plus-395-128gb-llama-cpp-tp1
schema version
local-ai-registry/v1

metrics

concurrency
1
inference engine version
llama.cpp
latest point at
2026-06-07T19:13:33.476Z
max context tokens
500,000
peak generation tps
137.91
peak prompt tps
772
point count
1

rows

concurrencycontext tokensdecode tok sdecode tok s per streamoutput tokenspeak vram gbprefill tok ssamples
1500,000137.91137.9150Unknown7721
Provenance & metadata (1)