Model · Language model
gpt-oss-120b
A developer-published mixture-of-experts language-model release whose downloadable checkpoints are governed by the unmodified Apache-2.0 license with a short accompanying usage statement.
- Developer
- OpenAI
- Distribution
- Downloadable weights
- Artifact licensing
- Open-licensed artifacts
- License scope
- Reviewed release artifacts and mandatory terms; not an assessment of training code or training data openness.
- Licenses
- Apache License 2.0
- Boundary
- This record represents the openai/gpt-oss-120b weight release at Hugging Face commit b5c939de8f754692c1647ca79fbf85e8c1e70f8a, last modified 2025-08-26. The gpt-oss-20b sibling, downstream quantizations and fine-tunes, OpenAI's managed API services, and independent hosting providers remain separate boundaries.
Risk statements
“High capability” · Below threshold
We ran scalable capability evaluations on gpt-oss-120b, and confirmed that the default model does not reach our indicative thresholds for High capability in any of the three Tracked Categories of our Preparedness Framework (Biological and Chemical capability, Cyber capability, and AI Self-Improvement).
- Risk areas
- Biological or chemical · Cyber · Autonomy
- Covers
- The model itself
Read the developer's statement ↗ · 2026-10-09 · Research confidence: Medium
OpenAI states that this release did not reach “High capability” in biological or chemical, cyber, and autonomy capability. The statement covers the model itself. This is the developer's own statement, not an Atlas risk rating.
Open in the directory → Open official model page ↗
Editorial ratings appear in the directory beside the profile they belong to and are never compared across collections.