Llama Guard 4 12B on NVIDIA

NVIDIA API Catalog (build.nvidia.com, NIM) Free API tier

The catch: Trial terms allow internal testing and evaluation only, not production, and let NVIDIA use your prompts and outputs to improve its products, including AI models. Free endpoints are retired often: DeepSeek V4, Qwen3.5, MiniMax, Mistral, GLM-5.1 and gpt-oss-120b lost theirs between July and September 2026.

Access	Free API tier
Free limits	Free Endpoint on build.nvidia.com for NVIDIA Developer Program members: up to 40 requests a minute and 10,000 a day (NVIDIA says limits may vary by model and other users' traffic can cause throttling).
Modality	text, image
Credit card	Not required
Commercial use	Not allowed
Trains on your data	Yes, on the free tier
Licence or terms	NVIDIA API Trial Terms of Service (v. September 19, 2025), plus each model's own licence
Context window	163,840 tokens
Model id	`meta/llama-guard-4-12b`
Base URL	`https://integrate.api.nvidia.com/v1`
Last checked	8 Oct 2026

How to use Llama Guard 4 12B

It speaks the OpenAI API at https://integrate.api.nvidia.com/v1, so the OpenAI SDK you already use works once you change the base URL and key.

Quickstart

curl https://integrate.api.nvidia.com/v1/chat/completions -H "Authorization: Bearer $NVIDIA_API_KEY" -H "Content-Type: application/json" -d '{"model":"meta/llama-guard-4-12b","messages":[{"role":"user","content":"Say OK"}]}'

Paste into Claude Code

I want to use the free AI model "Llama Guard 4 12B" from NVIDIA (model id: meta/llama-guard-4-12b) in my project.
Please:
1. Check the provider docs and give me the exact steps to get a free API key.
2. Give me a minimal working example that reads the key from an environment variable.
3. Tell me the real free limits and any catch (commercial use, card, data use). Listed limits: Up to 40 RPM · 10,000 requests/day.
4. If the same provider has a more generous free option, mention it.

Provider docs Get a free key Compare all 507 free options

Before you ship on it

Trial terms allow internal testing and evaluation only, not production, and let NVIDIA use your prompts and outputs to improve its products, including AI models. Free endpoints are retired often: DeepSeek V4, Qwen3.5, MiniMax, Mistral, GLM-5.1 and gpt-oss-120b lost theirs between July and September 2026.
Prompts on the free tier may be used to train models. Keep customer data off it.
NVIDIA API Trial Terms of Service (v. September 19, 2025) s.1.2: access "for limited trial purposes only and without use of the API Service or Generated Content in production"; s.1.4: without a paid Subscription "you…
Free tiers are shared capacity: treat the limits as a ceiling, not a promise, and keep a paid fallback.

Questions

Is Llama Guard 4 12B on NVIDIA free?

Yes. NVIDIA offers it as free api tier, with these limits: Up to 40 RPM · 10,000 requests/day. No credit card is needed.

Can I use Llama Guard 4 12B on NVIDIA commercially?

No: the free access is for testing or non-commercial use. A paid plan is needed for production.

How do I start with Llama Guard 4 12B on NVIDIA?

It speaks the OpenAI API at https://integrate.api.nvidia.com/v1, so the OpenAI SDK you already use works once you change the base URL and key.

Source: build.nvidia.com/meta/llama-guard-4-12b, checked 8 Oct 2026.