NVIDIA drops Nemotron 3.5 Content Safety

// 2h agoMODEL RELEASE

NVIDIA drops Nemotron 3.5 Content Safety

NVIDIA released Nemotron 3.5 Content Safety, a 4-billion parameter guardrail model fine-tuned from Google's Gemma-3-4B to moderate LLM inputs and outputs. It classifies text and images across 23 safety categories in 12 languages and features a chain-of-thought "THINK Mode" to explain safety decisions.

// ANALYSIS

AI safety is evolving from a static blocklist into a programmable policy runtime.

–Built-in reasoning (THINK Mode) addresses the black-box problem of LLM moderation, making decisions auditable and easier to debug.
–The 4B parameter size allows the model to run efficiently as a sidecar alongside main application LLMs without causing unacceptable latency.
–Relying on the Aegis v2 taxonomy provides a standardized framework that fits enterprise safety compliance workflows.

// TAGS

nvidianemotronsafetycontent-safetyguardrailsgemma-3model-release

DISCOVERED

2h ago

2026-06-06

PUBLISHED

2h ago

2026-06-06

RELEVANCE

8/ 10

AUTHOR

dball1126

// KEEP READING

More AI developer news from the feed

EXPLORE FULL FEED

MODEL1h ago

OpenRouter offers NVIDIA Nemotron 3 Ultra free

OpenRouter has integrated NVIDIA's Nemotron 3 Ultra model, offering it to users at zero cost. The massive Mixture-of-Experts model features a hybrid Transformer-Mamba architecture, 55B active parameters, and a 1M token context window for complex reasoning and agentic workflows.

UPDATE1h ago

OpenAI Codex adds remote mobile steering

OpenAI Codex has evolved into an agentic coding partner that runs locally on developer machines and integrates directly with the ChatGPT mobile app. This integration allows developers to remotely monitor, approve, and steer ongoing, long-running coding tasks from their mobile devices while on the go.

OPEN SOURCE1h ago

MemPalace introduces an open-source, local-first AI memory system designed to combat AI amnesia by storing conversation histories verbatim.

MemPalace is an open-source, local-first memory system designed to provide persistent, queryable memory for AI agents to solve the problem of context loss between chat sessions. Unlike other memory solutions that use LLMs to summarize conversations and extract key facts, MemPalace stores full conversation histories verbatim. It uses ChromaDB for semantic vector searches and SQLite for data infrastructure, organizing information into a hierarchical structure consisting of wings, rooms, and drawers. This allows AI coding tools and agents to perform scoped semantic searches to retrieve past decisions, preferences, and architectural choices securely on a user's local machine.

NVIDIA drops Nemotron 3.5 Content Safety