Llama Guard 4 12B

StableContent SafetyMeta

Llama Guard 4 is a Llama 4 Scout-derived multimodal pretrained model, fine-tuned for content safety classification. Similar to previous versions, it can be used to classify content in both LLM...

P50 32msP95 100ms$0.22 / 1M tokens164K ctx

Overview

Llama Guard 4 is a Llama 4 Scout-derived multimodal pretrained model, fine-tuned for content safety classification. Similar to previous versions, it can be used to classify content in both LLM... Llama Guard 4 12B is served through the PARROTSIGHT unified API at Meta's list price, with the same authentication, metering and regional pinning (MY / SG / ID) as every native model. No separate provider account is required.

164K token context · multimodal input
OpenAI-compatible endpoint — migrate with a two-line change
Streaming (SSE) supported for all text outputs
Billed at provider list price, metered per token

Capabilities

Detection quality86

False-positive control77

Speed80

Value89

API

https://api.parrotsight.com/v1/vision/analyze

bash

curl -X POST https://api.parrotsight.com/v1/vision/analyze \
  -H "Authorization: Bearer $PS_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{"model":"meta-llama/llama-guard-4-12b","image":"data:image/jpeg;base64,<BASE64_ENCODED_FRAME>","tasks":["scene","objects","caption"]}'

Version history

v2025.4Latest2025-04-30
Added to the PARROTSIGHT catalogue at Meta list pricing.

More models

View all models

nvidia/nemotron-3.5-content-safetyStable

Nemotron 3.5 Content Safety

NVIDIA Nemotron 3.5 Content Safety is a compact 4B-parameter multimodal guardrail model from NVIDIA, fine-tuned from Google Gemma-3-4B. It moderates both inputs to and responses from LLMs and VLMs, accepting...

$0.25 / 1M tokens

ps-guard-maxStable

SightGuard Max

Threat intelligence and anomaly detection at network scale

$2.80 / 1k calls