AxForge
Sign in Get started

Model reference · open weights

Llama-Guard-3

LLMs meta-llama Text gen 4 builds Open, with conditions 201k dl/mo

Llama-Guard-3 is an open-weight language model from Meta. Llama-Guard-3-8B (BF16) weighs 16.1 GB; the smallest configuration that runs it is RTX 4090 24 GB.

  • Llama-Guard-3 is an 8.0B parameter text-generation model developed by Meta for content safety classification.
  • It evaluates LLM inputs and responses against the MLCommons hazard taxonomy in eight languages: English, German, French, Italian, Portuguese, Hindi, Spanish, and Thai.
  • The model is released under the llama3.1 license.

Summary of the meta-llama/Llama-Guard-3-8B model card, 2026-10-01

Hugging Face ↗GitHub ↗Paper ↗

What it is

Released byMeta
Released2024-07-22
Parameters8.0B
VRAM16.1 GB for the weights

What it runs on

Memory and cards for Llama-Guard-3-8B (BF16)

16.1 GBweights, file size
762 MBruntime overhead, at least

How much memory each request adds isn't estimated yet for this architecture. The weights need at least the cards below, plus room for the context.

CardWeights alone
RTX 3060 12 GB … RTX 4060 Ti 16 GB
2 smaller cards		does not fit
RTX 3090 24 GBfits
RTX 4090 24 GBfits
RTX 5090 32 GBfits
L40S 48 GBfits
A100 80 GBfits
H100 80 GBfits
RTX PRO 6000 Blackwell 96 GBfits
DGX Spark (GB10) 128 GB unifiedfits
H200 141 GBfits
B200 180 GBfits

Builds

Sizes, precisions & builds

BuildParamsPrecisionWeightsSmallest setup
Llama-Guard-3-8B (above) ↗ 8.0BBF16 16.1 GBRTX 4090 24 GB
Llama-Guard-3-1B ↗ 1.5BBF16 3.0 GBRTX 3060 12 GB
Llama-Guard-3-8B-INT8 ↗ 8.0BINT8 9.1 GBRTX 3060 12 GB
Llama-Guard-3-1B-INT4 ↗ —INT4 ——

How it works

How language models work

Your prompttext / messagesTransformerattention over tokensNext-token loopgenerate + streamResponsetext · tool callsA language model reads your tokens and predicts the next one, again and again, streaming the reply back.
© 2026 AxForge · EU-hosted AI infrastructure Pricing Docs Trust Privacy Terms