models · poolside

Laguna-S-2.1

Catalogued from local.ai local-inference benchmark index at Quarantine — unverified, pending attestation.

What this registry has checked

Quarantine. We hold the publisher’s own published hash for these weights and have recorded it against this entry. We have not downloaded and hashed the weights ourselves, so the hash is the publisher’s claim, not our measurement.

Method	hf-published-sha256

Re-check this yourself → Source → Full record →

What is recorded

Publisher	poolside
Origin	Not recorded. The harvest wrote its own source name into this field, which tells us nothing about jurisdiction.
Licence	OpenMDW-1.1
Parameters	117,561,977,600 (117.6B)
Context window	262,144 tokens
Modality	text->text
Architecture	laguna
Layers	48
Hidden size	3072
Attention	GQA (48 heads, 8 KV)
Native precision	BF16
Downloads reported upstream	64,991

Measured on real hardware

None of these runs are ours. Each figure is the reporter’s, linked to their own evidence, under their own trust label. The number shown is single-stream decode — one user, concurrency 1 — because aggregate throughput across many users is a different quantity and reads as far faster than anyone will see.

Reported to local-ai-registry

Hardware	Engine	Build measured	Decode, 1 stream	Label	Reported by
Radeon AI PRO R9700 ×3 devices	llama.cpp	Q4_0_ROCMFP4_COHERENT	52.3 tok/s @ 2,048	candidate	localmaxxing
Radeon AI PRO R9700 ×3 devices	llama.cpp	Q4_0_ROCMFP4_STRIX_LEAN	52.3 tok/s @ 2,048	candidate	localmaxxing
Radeon AI PRO R9700 ×3 devices	llama.cpp	Q4_0_ROCMFP4_FAST	52.2 tok/s @ 2,048	candidate	localmaxxing
Measured outside 2K–32K context — decode speed depends heavily on context, so these are not comparable with the rows above
Radeon AI PRO R9700 ×3 devices	llama.cpp	UD-Q4_K_XL	51.6 tok/s @ 131,072	candidate	localmaxxing
Ryzen AI Max+ 395	llama.cpp	Q4_0_ROCMFP4_STRIX_LEAN	41.7 tok/s @ 262,144	candidate	localmaxxing
Ryzen AI Max+ 395	llama.cpp	Q4_0_ROCMFP4_FAST	40.7 tok/s @ 820	candidate	localmaxxing
Ryzen AI Max+ 395	llama.cpp	UD-Q4_K_XL	29.9 tok/s @ 820	candidate	localmaxxing

From local-ai-registry (MIT), commit 124959d: 8 runs across 3 machines, 0 validated — model revision and runtime pinned, launch accepted — and 8 candidate, their label for useful evidence without a reproducible-launch promise.

Measured by local.ai

Hardware	Build measured	Decode @ 8K	Memory @ 8K	Decode mode
NVIDIA RTX PRO 6000 Blackwell 96GB	Q4_K_M	117 tok/s	80.1 GB	ordinary
MacBook Pro M5 Max 128GB · 40-core GPU	Q4_K_M	48.3 tok/s	87.0 GB	ordinary
Mac Studio M3 Ultra 96GB · 80-core GPU	Q4_K_M	47.4 tok/s	88.7 GB	ordinary
Mac Studio M3 Ultra 96GB · 60-core GPU	Q4_K_M	47.1 tok/s	78.1 GB	ordinary
NVIDIA DGX Spark 128GB	Q4_K_M	20.1 tok/s	86.6 GB	ordinary
NVIDIA DGX Spark 128GB	NVFP4	18.5 tok/s	73.3 GB	ordinary

Measured by local.ai, not by us: 6 runs. Each names its engine pinned by image digest and the exact serve command, in the full record.

Elsewhere in the registry

All models → Browse with filters → What will run on my hardware? →

Registry record as of 2026-09-26. Identity, licence and parameter count are properties of a release and do not move; tier can, and the verify link above re-checks it against the live registry rather than this page.