LLM Hosting Pro
← All model families

gpt-oss

OpenAI's open-weight reasoning models, released under Apache 2.0.

ReasoningTool use

Overview

gpt-oss is OpenAI's family of open-weight reasoning models, released in 2025 in two sizes: gpt-oss-120b and gpt-oss-20b.

Both use the Apache 2.0 licence and support tool use and adjustable reasoning effort. The smaller model runs on a single high-memory consumer GPU.

Models in this family

ModelReleasedParametersContext
gpt-oss-20b5 Aug 202521B total, 3.6B active (mixture of experts)128K tokens
gpt-oss-120b5 Aug 2025117B total, 5.1B active (mixture of experts)128K tokens

Where to host gpt-oss models

All hosts →
LLM hostModelsHosting plans
Amazon AWSgpt-oss-20b, gpt-oss-120bAmazon EC2, Amazon Bedrock, Amazon SageMaker
Google Cloudgpt-oss-20b, gpt-oss-120bGoogle Kubernetes Engine (GKE), Compute Engine, Model Garden
Microsoft Azuregpt-oss-20b, gpt-oss-120bAzure AI Foundry, Azure Machine Learning, Azure Kubernetes Service (AKS), Azure GPU Virtual Machines