NVIDIA: Nemotron 3.5 Content Safety (free)
🧠 AI Modelnvidia
Free 4B multimodal guardrail for LLM/VLM content safety.
NVIDIA Nemotron 3.5 Content Safety is a specialized guardrail model designed to ensure safe interactions with large language models (LLMs) and vision-language models (VLMs). Fine-tuned from Google's Gemma-3-4B, it leverages a compact 4B parameter architecture for efficiency. The model supports multimodal inputs (text and images) and outputs safety classifications. It features a context length of 128,000 tokens, enabling thorough analysis of long conversations. Key capabilities include reasoning over safety policies, adjustable temperature and top-p sampling, and seed-based reproducibility. Hosted on OpenRouter as a free offering, it provides a cost-effective solution for content moderation and safety filtering in AI applications.
💡Highlights
- ├─4B parameters, fine-tuned from Gemma-3-4B
- ├─Multimodal: text and image input
- └─128K context length, free on OpenRouter
🎯For
- ├─AI developers
- ├─content moderators
- └─enterprise security teams