Context window
131,072 tokens
Checked 19 Oct 2026https://developers.openai.com/api/docs/models/gpt-oss-20b
LLM · OpenAI
gpt-oss-20b
OpenAI's model page describes gpt-oss-20b as an open-weight model with 21B parameters and 3.6B active parameters, a 131,072 token context window, and an Apache 2.0 license. It links a Hugging Face download. The endpoint table lists Responses and Batch as supported. Build, Launch, and Grow are listed at 0 RPM. The page does not print a token price.
Checked 19 Oct 2026https://developers.openai.com/api/docs/models/gpt-oss-20b
Input price
Not published
Checked 19 Oct 2026https://developers.openai.com/api/docs/models/gpt-oss-20b
Output price
Not published
Checked 19 Oct 2026https://developers.openai.com/api/docs/models/gpt-oss-20b
API deployment
Yes
Checked 19 Oct 2026https://developers.openai.com/api/docs/models/gpt-oss-20b
Local / self-hosted
Yes
Checked 19 Oct 2026https://developers.openai.com/api/docs/models/gpt-oss-20b
VPC / cloud native
Not published
Checked 19 Oct 2026https://developers.openai.com/api/docs/models/gpt-oss-20b
License
Apache 2.0
Checked 19 Oct 2026https://developers.openai.com/api/docs/models/gpt-oss-20b
Data retention
Not published
Checked 19 Oct 2026https://developers.openai.com/api/docs/models/gpt-oss-20b
Parameters
3.6B active / 21B total
Checked 19 Oct 2026https://developers.openai.com/api/docs/models/gpt-oss-20b