Speed sweep

pg-019f8333-8fd4-77b3-9d04-de9d4c0a3d0f-sweep

pg-019f8333-8fd4-77b3-9d04-de9d4c0a3d0f-sweep

JSON API Back to speed sweep search

Record

Recipe: pg-mlx-community-qwen3-6-27b-optiq-4bit-4bit-apple-m4-max-128gb-mlx-acc5662c88
Measured: 2026-07-21T05:43:42.164Z
Points: 4

Measured speed

Concurrency	Context	Prefill	Decode	TTFT ms	Status
1	8,256	—	22.2	—	observed
1	32,832	—	20	—	observed
1	65,600	—	17.6	—	observed
1	32,768	—	—	169,848.2	observed

Remaining fields

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

id: pg-019f8333-8fd4-77b3-9d04-de9d4c0a3d0f-sweep
measured at: 2026-07-21T05:43:42.164Z
recipe id: pg-mlx-community-qwen3-6-27b-optiq-4bit-4bit-apple-m4-max-128gb-mlx-acc5662c88
schema version: local-ai-registry/v1

metrics

base memory bytes: 19457777831
base memory context tokens: 192
concurrency: 1
decode32k context tokens: 32,832
decode32k tps: 19.962
decode8k context tokens: 8,256
decode8k tps: 22.22
decode max context tokens: 65,600
decode max context tps: 17.606
decode mode: non-mtp
inference engine version: 0.31.3
latest point at: 2026-07-21T05:43:42.164Z
max context tokens: 65,600
max prompt tokens: 65,536
memory8k bytes: 23153319166
memory8k context tokens: 8,256
memory max context bytes: 36241057030
memory max context tokens: 65,600
peak generation tps: 19.962
peak memory bytes: 36241057030
peak prompt tps: 192.925
point count: 56
ttft32k cached prompt tokens: 28,736
ttft32k context tokens: 32,768
ttft32k seconds: 169.848

rows

concurrency	context tokens	decode tok s	decode tok s per stream	output tokens	peak vram gb	prefill tok s	samples
1	8,256	22.22	22.22	Unknown	21.563	Unknown	1
1	32,832	19.962	19.962	Unknown	33.752	Unknown	1
1	65,600	17.606	17.606	Unknown	33.752	Unknown	1
1	32,768	Unknown	Unknown	Unknown	Unknown	Unknown	1

Provenance & metadata (1)

source

paths: publication:pg-20260827T060320709Z, run:019f8333-8fd4-77b3-9d04-de9d4c0a3d0f
repository: local.ai ↗