Model dossier

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...

ProviderOpenAI Start comparison

Context131,072

Max output65,536

Pricing viewTVP price

Identity

OpenAI: gpt-oss-safeguard-20b

Model ID: openai/gpt-oss-safeguard-20b
Canonical slug: openai/gpt-oss-safeguard-20b
Knowledge cutoff: n/a
Created: 10/29/2025

Capabilities

texttoolsstructured

Modalities

text->texttext

Capacity

Context131,072

Max output65,536

TokenizerGPT