gpt-oss-120b at a glance
The Apache 2.0 gpt-oss models are mixture-of-experts reasoning models: the 120b nears o4-mini on core reasoning on a single 80 GB GPU, and the 20b matches o3-mini and runs in 16 GB of memory.
What changed with gpt-oss-120b
On August 5, 2025, OpenAI released gpt-oss-120b and gpt-oss-20b, open-weight language models under the Apache 2.0 license. Both are mixture-of-experts transformers: the 120b has 117 billion total and 5.1 billion active parameters, the 20b has 21 billion total and 3.6 billion active, and both support 128K context.
OpenAI said gpt-oss-120b reaches near-parity with o4-mini on core reasoning benchmarks while running on a single 80 GB GPU, and gpt-oss-20b performs similarly to o3-mini and can run on devices with 16 GB of memory. They support low, medium and high reasoning effort, full chain-of-thought access, tool use such as web search and Python, and the Responses API format.
gpt-oss-120b in numbers
Values are copied from the linked source on the date shown. They are not IA Compa measurements.
- Context windowas of · Source for Context window: openai.com (opens in a new tab)
- 128K tokens
How it compares
Other models in the same category that we have covered, using each model's latest recorded value. (Models)
Where to compare gpt-oss-120b
Independent leaderboards and official resources for comparing this model yourself. Rankings change often; each site shows its own current data.
General leaderboards
Why gpt-oss-120b matters for businesses
For businesses, gpt-oss made it possible to run an OpenAI-grade reasoning model fully on your own hardware with a permissive license, a strong option for private deployments, regulated data and offline use.
Sources for this story
- Introducing gpt-oss (opens in a new tab)Official
OpenAI · openai.com
- openai/gpt-oss-120b (opens in a new tab)Official
Hugging Face · huggingface.co
- gpt-oss-120b model page (opens in a new tab)Benchmark
Artificial Analysis · artificialanalysis.ai
Last verified: . IA Compa summarizes public announcements for business readers. We are not affiliated with the companies mentioned. Always confirm details, pricing, and availability with the official source.
- #llm
- #openai
- #open-weights
- #reasoning
- #self-hosting
