Speed sweep

pg-019f2c34-d4fa-7c69-9c61-944738623230-sweep

pg-019f2c34-d4fa-7c69-9c61-944738623230-sweep

JSON API Back to speed sweep search

Record

Recipe: pg-mlx-community-gpt-oss-20b-mxfp4-q8-mxfp4-q8-apple-m4-max-36gb-mlx-4f5e5aa70f
Measured: 2026-08-09T16:05:24.472Z
Points: 4

Measured speed

Concurrency	Context	Prefill	Decode	TTFT ms	Status
1	2,111	—	86.1	—	observed
1	32,831	—	57.3	—	observed
1	131,007	—	29.3	—	observed
1	32,767	—	—	33,376.7	observed

Remaining fields

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

id: pg-019f2c34-d4fa-7c69-9c61-944738623230-sweep
measured at: 2026-08-09T16:05:24.472Z
recipe id: pg-mlx-community-gpt-oss-20b-mxfp4-q8-mxfp4-q8-apple-m4-max-36gb-mlx-4f5e5aa70f
schema version: local-ai-registry/v1

metrics

base memory bytes: 12284006990
base memory context tokens: 191
concurrency: 1
decode32k context tokens: 32,831
decode32k tps: 57.271
decode8k context tokens: 2,111
decode8k tps: 86.118
decode max context tokens: 131,007
decode max context tps: 29.258
decode mode: non-mtp
inference engine version: 0.31.3
latest point at: 2026-08-09T16:05:24.472Z
max context tokens: 131,007
max prompt tokens: 130,943
memory8k bytes: 13005211082
memory8k context tokens: 2,111
memory max context bytes: 18596182436
memory max context tokens: 131,007
peak generation tps: 57.271
peak memory bytes: 18596182436
peak prompt tps: 981.733
point count: 16
ttft32k cached prompt tokens: 0
ttft32k context tokens: 32,767
ttft32k seconds: 33.377

rows

concurrency	context tokens	decode tok s	decode tok s per stream	output tokens	peak vram gb	prefill tok s	samples
1	2,111	86.118	86.118	Unknown	12.112	Unknown	1
1	32,831	57.271	57.271	Unknown	17.319	Unknown	1
1	131,007	29.258	29.258	Unknown	17.319	Unknown	1
1	32,767	Unknown	Unknown	Unknown	Unknown	Unknown	1

Provenance & metadata (1)

source

paths: publication:pg-20260827T060320709Z, run:019f2c34-d4fa-7c69-9c61-944738623230
repository: local.ai ↗