We’re launching Lyria 3.5 in Google Flow Music, with advances across musicality, lyrics, vocals, and creative control
deepmind.google·6d ago
AI Engineering
OpenAI Blog·Aug 5·5 min read
gpt-oss-120b & gpt-oss-20b Model Card
TL;DR
There was a need for open-weight reasoning models to enhance accessibility and collaboration in AI research. OpenAI developed gpt-oss-120b and gpt-oss-20b, two models with 120 billion and 20 billion parameters, respectively.
✦ Why It Matters
Engineers and researchers can leverage these open-weight models to enhance their AI projects and foster innovation.
Key Takeaways
- Explore gpt-oss models for your AI projects.
- Utilize the Apache 2.0 license for collaboration.
- Experiment with different parameter sizes for efficiency.
- Contribute to the community by sharing findings.
Full Summary
OpenAI identified a gap in the availability of large-scale reasoning models that researchers and developers could freely use and modify. In response, they created gpt-oss-120b and gpt-oss-20b, which are open-weight models designed to facilitate advanced reasoning tasks.
The gpt-oss-120b model contains 120 billion parameters, while the gpt-oss-20b model has 20 billion parameters, allowing for varying levels of computational resource requirements. Both models are released under the Apache 2.0 license, which encourages collaboration and innovation in AI research.
By making these models accessible, OpenAI aims to foster a community-driven approach to AI development. The implications of this release include increased opportunities for experimentation and the potential for new applications in various fields.