models · meta-llama

Llama-3.2-3B-Instruct

Compact 3B instruct model for on-device and edge deployment with 128K context. Distilled from the larger Llama 3.1 models.

Compare Verify

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.

Weight file SHA-256	9b1ab57fbb64db069c8d9839fbfb1949c89281dd92703605f8e765bbc317c260
Method	self-download-failed

Re-check this yourself → Source → Full record →

What is recorded

Publisher	meta-llama
Origin	United States
Licence	Llama-3.2
Publisher’s own licence statement	llama3.2
Parameters	3,212,749,824 (3.21B)
Context window	128K
Modality	text->text
Native precision	BF16
Pulls recorded here	224,014
Downloads reported upstream	1,690,943

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
GeForce RTX 3060	llama.cpp	Q4_K_M	128 tok/s @ 512	candidate	localmaxxing
Apple M1 Max 64GB	llama.cpp	BF16	50.9 tok/s @ 640	candidate	localmaxxing

From local-ai-registry (MIT), commit 124959d: 2 runs across 2 machines, 0 validated — model revision and runtime pinned, launch accepted — and 2 candidate, their label for useful evidence without a reproducible-launch promise.

Elsewhere in the registry

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

Registry record as of 2026-06-12. 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.