OpenAI Releases Open-Weight Safety Classifiers gpt-oss-safeguard
OpenAI's gpt-oss-safeguard-120b and -20b classify content against developer-written policies at inference time, beating gpt-5-thinking on multi-policy accuracy under Apache 2.0.
Topic
Topic
OpenAI's gpt-oss-safeguard-120b and -20b classify content against developer-written policies at inference time, beating gpt-5-thinking on multi-policy accuracy under Apache 2.0.