OpenAI rolls out GPT‑OSS model: What is it, who can use, and all other details

OpenAI

OpenAI has debuted its first open‑weight language models in over six years. The GPT‑OSS lineup — the first release since GPT‑2 —marks a strategic shift toward openness in response to increased competition from China’s DeepSeek and others.

In a post on X, OpenAI CEO, Sam Altman said, “We’re excited to make this model, the result of billions of dollars of research, available to the world to get AI into the hands of the most people possible. We believe far more good than bad will come from it.”

Two model sizes for different use cases

GPT‑OSS comes in two variants: gpt-oss-120b (~120 billion parameters) for high-reasoning, production workflows, and gpt-oss-20b (~21 billion parameters) optimised for lower latency and even consumer hardware with 16 GB RAM. This means these models can run on your desktop or laptop as well.

GPT‑OSS supports “chain-of-thought” reasoning, function calling (web search, code execution), and a configurable reasoning mode (low/medium/high) to trade depth for speed. This makes it ideal for agentic workflows and developer customization  .

Why this matters

OpenAI’s release directly responds to global momentum behind open models from DeepSeek, Alibaba, Meta, Mistral, and others. It aims to retain developer mindshare and promote US‑based innovation rooted in democratic values, as per the company. By lowering barriers to access and enabling customization, GPT‑OSS opens new doors for startups, researchers, and institutions that require local, transparent AI, claims OpenAI.

Altman acknowledged past reluctance — citing safety concerns — but says the release aligns with the company’s mission to build AGI that benefits all of humanity. “OpenAI’s mission is to ensure AGI that benefits all of humanity. To that end, we are excited for the world to be building on an open AI stack created in the United States, based on democratic values, available for free to all and for wide benefit,” added Altman in the post.