OpenAI has unveiled two advanced open-weight language models, gpt-oss-120b and gpt-oss-20b, under the Apache 2.0 license. These models are engineered for robust reasoning and agentic tasks, offering developers the flexibility to customize and deploy them across various applications. The gpt-oss-120b, with 120 billion parameters, is optimized for high-end hardware, delivering performance comparable to OpenAI's proprietary o4-mini model. In contrast, the gpt-oss-20b, featuring 20 billion parameters, is designed for efficient operation on consumer-grade devices, making it accessible for a broader range of users. Both models excel in tasks such as coding, mathematical reasoning, and health-related queries, and are optimized for efficient deployment on consumer hardware. This release marks a significant step in OpenAI's commitment to democratizing AI technology, providing the community with powerful tools for innovation and development.