Speed sweep

pg-019f2c34-e7d3-7cfb-ab4f-7d26b238365c-sweep

pg-019f2c34-e7d3-7cfb-ab4f-7d26b238365c-sweep

JSON API Back to speed sweep search

Record

Recipe: pg-mlx-community-qwen3-6-35b-a3b-4bit-4bit-apple-m5-pro-64gb-mlx-15632efe76
Measured: 2026-07-04T08:18:12.559Z
Points: 3

Measured speed

Concurrency	Context	Prefill	Decode	TTFT ms	Status
1	8,256	—	79.4	—	observed
1	32,832	—	66.3	24,208.9	observed
1	258,111	—	26.1	—	observed

Remaining fields

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

id: pg-019f2c34-e7d3-7cfb-ab4f-7d26b238365c-sweep
measured at: 2026-07-04T08:18:12.559Z
recipe id: pg-mlx-community-qwen3-6-35b-a3b-4bit-4bit-apple-m5-pro-64gb-mlx-15632efe76
schema version: local-ai-registry/v1

metrics

base memory bytes: 21290696714
base memory context tokens: 192
concurrency: 1
decode32k context tokens: 32,832
decode32k tps: 66.254
decode8k context tokens: 8,256
decode8k tps: 79.395
decode max context tokens: 258,111
decode max context tps: 26.139
decode mode: non-mtp
inference engine version: unknown
latest point at: 2026-07-04T08:18:12.559Z
max context tokens: 258,111
max prompt tokens: 258,047
memory8k bytes: 21975767070
memory8k context tokens: 8,256
memory max context bytes: 48366876950
memory max context tokens: 258,111
peak generation tps: 66.254
peak memory bytes: 48366876950
peak prompt tps: 1,353.554
point count: 223
ttft32k cached prompt tokens: 28,736
ttft32k context tokens: 32,768
ttft32k seconds: 24.209

rows

concurrency	context tokens	decode tok s	decode tok s per stream	output tokens	peak vram gb	prefill tok s	samples
1	8,256	79.395	Unknown	Unknown	20.467	Unknown	1
1	32,832	66.254	Unknown	Unknown	45.045	Unknown	1
1	258,111	26.139	Unknown	Unknown	45.045	Unknown	1

Provenance & metadata (1)

source

paths: publication:pg-20260827T060320709Z, run:019f2c34-e7d3-7cfb-ab4f-7d26b238365c
repository: local.ai ↗