Local AI Registry

Recipe

unsloth-gemma-4-12b-it-gguf--q4-k-m-rtx-3060-ti-8gb-ollama-tp1

unsloth-gemma-4-12b-it-gguf--q4-k-m-rtx-3060-ti-8gb-ollama-tp1
JSON APIFind compatible recipes

Measured ollama run of gemma-4-12b-it on a rented Rtx 3060 Ti 8Gb via the registry lane campaign.

Record

Status
candidate
Source
lane-experiment
Engine
ollama
Engine version
0.33.2
Accelerators
1
Tensor parallel
1
Context tokens
16,384
chat
unknown
reasoning
unknown
tools
unknown
vision
unknown

Hugging Face model card

Identity

https://huggingface.co/unsloth/gemma-4-12b-it-GGUF
Repository
unsloth/gemma-4-12b-it-GGUF
Status
known
Link type
Exact Hub repository

Public Hugging Face repository confirmed by the Hub API.

Candidate: useful compatibility or speed evidence. The registry does not offer Run until promotion requirements are met.

Observed configuration

ollama

Evidence only · candidate · reference

Candidate evidence — not a Run contract

Source
https://console.vast.ai/instances/?label=ampere-lane

No tokenized launch fields. This is measured or documented compatibility, not a Docker launch.

Measured speed

ConcurrencyContextPrefillDecodeTTFT msStatusSweep
116,384———observedunsloth-gemma-4-12b-it-gguf--q4-k-m-rtx-3060-ti-8gb-ollama-tp1-sweep

Remaining fields

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

hardware count
1
hardware id
rtx-3060-ti-8gb
id
unsloth-gemma-4-12b-it-gguf--q4-k-m-rtx-3060-ti-8gb-ollama-tp1
model instance id
unsloth-gemma-4-12b-it-gguf--q4-k-m
recipe source
lane-experiment
schema version
local-ai-registry/v1
speed sweep ids
unsloth-gemma-4-12b-it-gguf--q4-k-m-rtx-3060-ti-8gb-ollama-tp1-sweep
status
candidate

capabilities

engine

name
ollama
version
0.33.2

serving

max context tokens
16,384
tensor parallel
1
Provenance & metadata (3)

facts

metadata

provenance

captured at
2026-09-02T16:45:30Z

sources

captured atkindurl
2026-09-02T16:45:30Zlane-experimentconsole.vast.ai/instances/ ↗