OpenAI: gpt-oss-safeguard-20b
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...
Everything else
| Max output | 66K |
|---|---|
| Input / 1M | $0.075 |
| Output / 1M | $0.3 |
| Cache read / 1M | $0.037 |
| Cache write / 1M | — |
| 1M in + 1M out | $0.375 |
| Takes | text |
| Returns | text |
| Tools | yes |
| Reasoning | always on |
| Open weights | yes |
| Released | 2025-10-29 |
| Knowledge cutoff | — |
| Retires | — |
| Catalog id | openai/gpt-oss-safeguard-20b |
Using it
In 00, this model is picked per agent — and per tier, so the model that answers your customers need not be the one that writes your code. Get 00.