title: “gpt-oss-20B”

gpt-oss-20B

OpenAI’s 20B parameter model, part of the [[concepts/gpt-4|gpt-oss]] family of open-weight language models, boasting state-of-the-art retrieval performance comparable to larger models. Key differentiators include efficient and accurate self-editing search capabilities for enhanced Retrieval-Augmented Generation (RAG).

Summary

• OpenAI released the gpt-oss family of open-weight language models, including gpt-oss-120b and gpt-oss-20b variants.
• Developed as the next evolution of RAG, it is a 20 billion parameter model trained specifically for RAG with self-editing search capabilities.
• Achieves retrieval performance comparable to larger models with significantly fewer parameters.

Related Concepts and Entities

• Retrieval-Augmented Generation (RAG)
• agentic-ai

2026 04 14 Jeredblu running LLM locally

• 2026-04-07 2026-04-07-Chroma-Context-1-Self-Editing-Search-Agent-for-Efficient-RAG ← Chroma Context 1 Self Editing Search Agent For Efficient Rag
• 2026-04-08 2026-04-08-Chroma-Context-1-Self-Editing-Search-Agent-for-Efficient-RAG ← Chroma Context 1 Self Editing Search Agent For Efficient Rag
• 2026-04-10 2026-04-10-Chroma-Context-1-Self-Editing-Search-Agent-for-Efficient-RAG ← Chroma Context 1 Self Editing Search Agent For Efficient Rag

Source Notes