**Fast and Affordable AI Inference For the World’s Latest Open Safety Model**

We’re excited to announce the availability of GPT-OSS-Safeguard-20B on GroqCloud, providing day zero, on‑demand access to OpenAI’s newest open‑source model running at over 1000 t/s. This first‑of‑its‑kind open‑weight reasoning model is purpose-built for safety‑classification workloads and lets you bring your own policy to production in minutes.

### **What is GPT‑OSS‑Safeguard‑20B?**

- Fine‑tuned from OpenAI’s GPT‑OSS-20B  
- Built for safety use cases  
- Safety‑first reasoning – trained to follow explicit, user‑provided policies and to explain its decisions.

In short, GPT-OSS-Safeguard-20B provides a reasoned classifier instead of a raw score, making debugging and compliance far easier.

### **Core Features for Trust & Safety Teams Enabled on GroqCloud**

- **Bring Your Own Policy:** Load any taxonomy, definition set, or threshold you own. The model will reason against it, letting you enforce exactly the rules that matter to your product.
- **Configurable Reasoning Effort (low / medium / high):** Trade latency for depth of analysis on a per‑request basis.
- **Full Reasoning Trace:** The model returns its chain‑of‑thought, giving you visibility into how a decision was reached (not for end‑users, but invaluable for audit logs and debugging).
- **Prompt Caching**: Re‑use the same policy prompt across millions of calls, significantly reducing latency and cost with up to 50% discount on cached tokens.
- **128k Token Context Window:** Process long documents, policy bundles, or multi‑turn conversations.

### **Delivering Price-Performance**

Groq's purpose-built stack is designed to deliver low cost per token for OpenAI’s new safeguard model while maintaining speed and accuracy. With **GPT-OSS-Safeguard-20B** available on GroqCloud, developers can run cutting-edge AI safety workloads tailored to their specific use case while keeping costs low and latency predictable.

[`openai/gpt-oss-safeguard-20b`](https://console.groq.com/playground?model=openai/gpt-oss-safeguard-20b) is currently running at 1000+ t/s on GroqCloud at the same price as the base GPT-OSS-20B model: $0.075 / M input tokens and $0.30 / M output tokens.

### **Example Use Cases**

GPT‑OSS‑Safeguard‑20B’s flexibility and bring your own policy capabilities make it a natural fit for many real‑world safety scenarios. Below are a few examples that illustrate how you can leverage the model in various applications.

- **Enterprise Document Scanning:** Define policies for confidential data (PII, PHI, financial info). Feed PDFs, emails, or contracts to the model; it flags violations and explains why.
- **AI‑Powered Chatbots**: Embed a safety policy in the system prompt of a conversational agent. The model blocks or rewrites unsafe user inputs before they reach the base model.
- **Automated Policy Auditing**: Feed existing moderation rules into the model as “policy‑in‑policy” and ask it to suggest gaps, contradictions, or improvements.
- **User‑Generated Content Platforms:** For forums, review sites, or marketplaces, run each new listing/comment through the model with a policy that captures fraud, prohibited items, or deceptive language.

### **Get Started with GPT-OSS-Safeguard-20B On GroqCloud**

Try [`openai/gpt-oss-safeguard-20b`](https://console.groq.com/playground?model=openai/gpt-oss-safeguard-20b) via the GroqCloud Developer Console, available in the Playground and the API.

To learn more about content moderation and best practices, check out our [developer documentation](https://console.groq.com/docs/content-moderation).