MediumCapability
OpenAI launches gpt-oss-safeguard: Bring-your-own-policy moderation in LM Studio
Action Required
Developers can now integrate a policy-aligned reasoning layer into their applications, improving safety and compliance without retraining.
AI Impact Summary
OpenAI has launched gpt-oss-safeguard, open safety reasoning models (120B and 20B) that allow users to bring their own policy moderation through LM Studio. This enables Trust & Safety teams to integrate a policy-aligned reasoning layer into existing moderation systems, offering faster iteration on policies without retraining. Developers can use the LM Studio SDK or OpenAI Responses API compatibility mode to call the model from their code, leveraging custom policy prompts.
OpenAI launches gpt-oss-safeguard: Bring-your-own-policy moderation in LM Studio
Affected Systems
LM Studio
- Date
- 29 Oct 2025
- Change type
- capability
- Severity
- medium