GPT-OSS is a family of open source language models from OpenAI, including the gpt-oss-120b and gpt-oss-20b versions with 117 billion and 210 billion parameters, respectively. The family is licensed under the Apache 2.0 license, which allows developers to download, modify, and deploy it for free.
Key features include:
- Efficient inference: using MXFP4 quantization, gpt-oss-120b runs on a single Nvidia H100 GPU, gpt-oss-20b adapts to a 16GB RAM device.
- Chained reasoning: supports low, medium, and high reasoning strengths that can be adjusted according to task requirements
- Tool invocation: integrated web search, Python code execution, file manipulation, etc.
- Structured Output: Uses proprietary Harmony response format for ease of debugging and use
- Long context support: 128k context length is natively supported.
This answer comes from the articleGPT-OSS: OpenAI's Open Source Big Model for Efficient ReasoningThe
May not be reproduced without permission:AI productivity tools " What is GPT-OSS? What are its main features?
English 
简体中文 
日本語 
Deutsch 
Português do Brasil