Speed sweep

pg-01a0348f-1985-7a8f-b670-0d6f7c2ded1f-sweep

pg-01a0348f-1985-7a8f-b670-0d6f7c2ded1f-sweep

JSON API Back to speed sweep search

Record

Recipe: pg-mlx-community-gpt-oss-120b-mxfp4-q8-mxfp4-q8-apple-m3-ultra-96gb-60c-mlx-7c85fffa26
Measured: 2026-08-24T16:16:28.534Z
Points: 4

Measured speed

Concurrency	Context	Prefill	Decode	TTFT ms	Status
1	8,256	—	61.1	—	observed
1	32,832	—	48.8	—	observed
1	65,600	—	38.7	—	observed
1	32,768	—	—	34,565.2	observed

Remaining fields

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

id: pg-01a0348f-1985-7a8f-b670-0d6f7c2ded1f-sweep
measured at: 2026-08-24T16:16:28.534Z
recipe id: pg-mlx-community-gpt-oss-120b-mxfp4-q8-mxfp4-q8-apple-m3-ultra-96gb-60c-mlx-7c85fffa26
schema version: local-ai-registry/v1

metrics

base memory bytes: 63553931184
base memory context tokens: 192
concurrency: 1
decode32k context tokens: 32,832
decode32k tps: 48.843
decode8k context tokens: 8,256
decode8k tps: 61.119
decode max context tokens: 65,600
decode max context tps: 38.74
decode mode: non-mtp
inference engine version: 0.31.3
latest point at: 2026-08-24T16:16:28.534Z
max context tokens: 130,947
max prompt tokens: 130,944
memory8k bytes: 64824936066
memory8k context tokens: 8,256
memory max context bytes: 68459219370
memory max context tokens: 65,600
peak generation tps: 48.843
peak memory bytes: 73151902426
peak prompt tps: 948.006
point count: 16
ttft32k cached prompt tokens: 0
ttft32k context tokens: 32,768
ttft32k seconds: 34.565

rows

concurrency	context tokens	decode tok s	decode tok s per stream	output tokens	peak vram gb	prefill tok s	samples
1	8,256	61.119	61.119	Unknown	60.373	Unknown	1
1	32,832	48.843	48.843	Unknown	68.128	Unknown	1
1	65,600	38.74	38.74	Unknown	63.758	Unknown	1
1	32,768	Unknown	Unknown	Unknown	Unknown	Unknown	1

Provenance & metadata (1)

source

paths: publication:pg-20260827T060320709Z, run:01a0348f-1985-7a8f-b670-0d6f7c2ded1f
repository: local.ai ↗