OpenAI has released two open-weight language models designed for advanced reasoning, with performance comparable to its smaller proprietary models (o3-mini and o4-mini). These models can run locally—one on a single GPU and the other even on personal laptops—offering flexibility for developers to analyze and fine-tune them without access to the original training data.
Open-weight models differ from open-source ones in that only the trained parameters are shared, not the full source code or training process. This marks OpenAI’s first open model release since GPT-2 in 2019.
The models—gpt-oss-120b and gpt-oss-20b—are particularly strong in coding, competitive math, and health queries, and were trained on a science- and math-focused text dataset. While they show strong performance, OpenAI has not yet released benchmark comparisons with rival models like DeepSeek-R1.
Credit: CGTN
You can listen to the programs organized by MI Radio on Google Podcast, Apple Podcast and Spotify.
You can download MI Radio application on Google Play Store and Apple App Store.
