Speed sweep

pg-019f7583-28c7-7987-b41a-508e2a8eb463-sweep

pg-019f7583-28c7-7987-b41a-508e2a8eb463-sweep

JSON API Back to speed sweep search

Record

Recipe: pg-mlx-community-qwen3-6-27b-optiq-4bit-4bit-apple-m5-pro-64gb-mlx-9dfa57a21c
Measured: 2026-07-18T13:55:57.723Z
Points: 4

Measured speed

Concurrency	Context	Prefill	Decode	TTFT ms	Status
1	8,256	—	14	—	observed
1	32,832	—	12.8	—	observed
1	65,600	—	11.4	—	observed
1	32,768	—	—	103,650.3	observed

Remaining fields

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

id: pg-019f7583-28c7-7987-b41a-508e2a8eb463-sweep
measured at: 2026-07-18T13:55:57.723Z
recipe id: pg-mlx-community-qwen3-6-27b-optiq-4bit-4bit-apple-m5-pro-64gb-mlx-9dfa57a21c
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: 12.814
decode8k context tokens: 8,256
decode8k tps: 14.05
decode max context tokens: 65,600
decode max context tps: 11.449
decode mode: non-mtp
inference engine version: 0.31.3
latest point at: 2026-07-18T13:55:57.723Z
max context tokens: 65,600
max prompt tokens: 65,536
memory8k bytes: 23153321214
memory8k context tokens: 8,256
memory max context bytes: 36241253638
memory max context tokens: 65,600
peak generation tps: 12.814
peak memory bytes: 36241253638
peak prompt tps: 316.14
point count: 56
ttft32k cached prompt tokens: 28,736
ttft32k context tokens: 32,768
ttft32k seconds: 103.65

rows

concurrency	context tokens	decode tok s	decode tok s per stream	output tokens	peak vram gb	prefill tok s	samples
1	8,256	14.05	14.05	Unknown	21.563	Unknown	1
1	32,832	12.814	12.814	Unknown	33.752	Unknown	1
1	65,600	11.449	11.449	Unknown	33.752	Unknown	1
1	32,768	Unknown	Unknown	Unknown	Unknown	Unknown	1

Provenance & metadata (1)

source

paths: publication:pg-20260827T060320709Z, run:019f7583-28c7-7987-b41a-508e2a8eb463
repository: local.ai ↗