Llama Guard 4 12B
meta/llama-guard-4-12b12B
Quality Index: 82
Safety guard model
Performance Benchmarks
Reasoning (Math)
75/100
Coding
60/100
Quality Index
82
Example Prompts
Llama: instruction following
Write a short product spec for an AI chat app. Include goals, non-goals, user stories, and success metrics.
General assistant
You are a helpful assistant. Answer clearly, concisely, and ask clarifying questions when needed.
Code review
Review this code for correctness, security, and performance. Provide prioritized fixes with explanations and patches.
Structured output
Return JSON with keys: summary, risks (array), recommendations (array). Do not include extra keys.
Pricing
Input$0.15 / 1M
Output$0.45 / 1M
Blended average (3:1)
$0.22 / 1M tokens
Technical Specs
Parameters12B
Context Window32,000
LicenseLlama 4 Community
Released2025-02-01