models · google

gemma-4-12B-it

any-to-any model harvested from Hugging Face (trending). Catalogued at Quarantine — unverified, pending attestation.

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.
Methodhf-published-sha256

What is recorded

Publishergoogle
OriginUnited States
Licenceapache-2.0
Parameters11,959,730,224 (12B)
Native precisionBF16
Pulls recorded here675,954
Downloads reported upstream2,958,440

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.

Measured by local.ai

HardwareBuild measuredDecode @ 8KMemory @ 8KDecode mode
NVIDIA GeForce RTX 5090 32GBq4km141 tok/s12.9 GBordinary
NVIDIA RTX PRO 6000 Blackwell 96GBq4km133 tok/s13.0 GBordinary
NVIDIA GeForce RTX 4090 24GBq4km90.1 tok/s12.8 GBordinary
NVIDIA RTX 6000 Ada 48GBq4km87.9 tok/s12.8 GBordinary
MacBook Pro M5 Max 128GB · 40-core GPUq4km53.8 tok/s14.6 GBordinary
Mac Studio M3 Ultra 96GB · 80-core GPUq4km52.2 tok/s14.6 GBordinary
Mac Studio M3 Ultra 96GB · 60-core GPUq4km52.1 tok/s14.6 GBordinary
MacBook Pro M4 Max 128GB · 40-core GPUq4km47.0 tok/s14.5 GBordinary

Measured by local.ai, not by us: 17 runs, 9 more not listed. Each names its engine pinned by image digest and the exact serve command, in the full record.

Elsewhere in the registry

Registry record as of 2026-09-01. 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.