Nemotron 3.5 Content Safety
Nemotron 3.5 Content Safety is a large language model made by NVIDIA, added to the catalog on 4 Jun 2026. It has a context window of 131,072 tokens and costs $0.20 per 1M input tokens and $0.20 per 1M output tokens. It accepts text and images and produces text. Typical uses: moderating prompts sent to language models, reviewing language model responses and moderating image-based inputs.
Price checked ยท source: OpenRouter
What Nemotron 3.5 Content Safety is
NVIDIA Nemotron 3.5 Content Safety is a multimodal guardrail model from NVIDIA for content moderation around language and vision-language models. It is intended to moderate both the inputs sent to those models and the responses they produce. It accepts text and image inputs and returns text. This makes it relevant to workflows where safety review needs to cover written material as well as image-based content. The catalog identifies its role as moderation but does not specify particular policy categories, labels, or a required response schema, so its output should be understood as text rather than a particular score or classification.
The model has a context window of 131,072 tokens and a maximum output of 117,964 tokens. That context capacity can suit moderation tasks involving long prompts, extended conversations, or substantial responses, as well as shorter pieces of content. Since it also accepts images, it can be considered for workflows that include visual inputs alongside text. The catalog describes it as a compact 4B-parameter model fine-tuned from Google Gemma-3-4B. These details describe its model family and input and output limits; the catalog does not specify response latency or moderation coverage. Nemotron 3.5 Content Safety is designed for review of content on both sides of interactions with LLMs and VLMs.
Key features
- Accepts text and image inputs
- Produces text outputs
- Has a 131,072-token context window
- Supports a maximum output of 117,964 tokens
- Designed to moderate inputs and responses from LLMs and VLMs
Typical uses
- Moderating prompts sent to language models
- Reviewing language model responses
- Moderating image-based inputs
- Safety review for vision-language model interactions
Nemotron 3.5 Content Safety pricing
US dollars, list price through OpenRouter.
| Item | Standard |
|---|---|
| Input, per 1M tokens | $0.20 |
| Output, per 1M tokens | $0.20 |
No price change since it was added on 4 Jun 2026.
| What a task costs | Cost |
|---|---|
| Answer a short question500 in / 300 out | $0.0002 |
| Summarise a 50-page PDF25,000 in / 800 out | $0.0052 |
| Review a 2,000-line source file30,000 in / 2,000 out | $0.0064 |
| Read a 300-page contract set150,000 in / 1,500 out | $0.030 |
Nemotron 3.5 Content Safety specifications
| Model id | nvidia/nemotron-3.5-content-safety |
|---|---|
| Version | nemotron-3.5-content-safety-20260604 |
| Context window | 131,072 tokensAbout 98,304 English words |
| Max output | 117,964 tokens |
| Knowledge cutoff | Not published |
| Added to catalog | 4 Jun 2026 |
| Input types | text, image |
|---|---|
| Output types | text |
| Tool calling | No |
| Structured outputs | No |
| Reasoning effort control | No |
| Content moderation | Not applied by provider |
Sample outputs
The same prompts are given to every model, and the unedited output is shown with the date it was produced. Upend does not publish leaderboard scores it did not run.
No output has been recorded for Nemotron 3.5 Content Safety yet.
Compare Nemotron 3.5 Content Safety
Side by side with other models: prices, context window, what each accepts and supports, and the same four tasks priced on both.