Speed sweep

pg-019f44e0-c606-7ecb-840b-c3c8aea300b0-sweep

pg-019f44e0-c606-7ecb-840b-c3c8aea300b0-sweep

JSON API Back to speed sweep search

Record

Recipe: pg-mlx-community-gemma-4-26b-a4b-it-4bit-4bit-apple-m4-max-128gb-mlx-4fb7dd353e
Measured: 2026-07-09T03:16:48.990Z
Points: 4

Measured speed

Concurrency	Context	Prefill	Decode	TTFT ms	Status
1	8,256	—	92.3	—	observed
1	32,832	—	72.8	—	observed
1	262,079	—	24.1	—	observed
1	32,768	—	—	32,740	observed

Remaining fields

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

id: pg-019f44e0-c606-7ecb-840b-c3c8aea300b0-sweep
measured at: 2026-07-09T03:16:48.990Z
recipe id: pg-mlx-community-gemma-4-26b-a4b-it-4bit-4bit-apple-m4-max-128gb-mlx-4fb7dd353e
schema version: local-ai-registry/v1

metrics

base memory bytes: 28463896942
base memory context tokens: 192
concurrency: 1
decode32k context tokens: 32,832
decode32k tps: 72.849
decode8k context tokens: 8,256
decode8k tps: 92.267
decode max context tokens: 262,079
decode max context tps: 24.146
decode mode: non-mtp
inference engine version: 0.31.3
latest point at: 2026-07-09T03:16:48.990Z
max context tokens: 262,079
max prompt tokens: 262,015
memory8k bytes: 28463896942
memory8k context tokens: 8,256
memory max context bytes: 43508255372
memory max context tokens: 262,079
peak generation tps: 72.849
peak memory bytes: 43508255372
peak prompt tps: 1,000.855
point count: 224
ttft32k cached prompt tokens: 28,736
ttft32k context tokens: 32,768
ttft32k seconds: 32.74

rows

concurrency	context tokens	decode tok s	decode tok s per stream	output tokens	peak vram gb	prefill tok s	samples
1	8,256	92.267	92.267	Unknown	26.509	Unknown	1
1	32,832	72.849	72.849	Unknown	40.52	Unknown	1
1	262,079	24.146	24.146	Unknown	40.52	Unknown	1
1	32,768	Unknown	Unknown	Unknown	Unknown	Unknown	1

Provenance & metadata (1)

source

paths: publication:pg-20260827T060320709Z, run:019f44e0-c606-7ecb-840b-c3c8aea300b0
repository: local.ai ↗