AI model · OpenAI
gpt-oss-safeguard-20b
Model ID: openai/gpt-oss-safeguard-20b
- Input
- $0.08 per 1M tokens
- Output
- $0.30 per 1M tokens
- Context
- 131K tokens at once
- Max output
- 66K tokens per reply
- Released
- 29 Oct 2025 OpenAI
What it can do
- YesReasoning
- YesTool calling
- NoImage input
- YesOpen weights
- NoFree to use
Takes text; gives text.
How its listing describes it
gpt-oss-safeguard-20b is a safety reasoning model from OpenAI built upon gpt-oss-20b. This open-weight, 21B-parameter Mixture-of-Experts (MoE) model offers lower latency for safety tasks like content classification, LLM filtering, and trust...
Sources
- OpenAI: gpt-oss-safeguard-20b on OpenRouter — read 2 Oct 2026