ModelWatch
← Catalog

gpt-oss-20b

gpt-oss-20b is an open-weight 21B parameter model released by OpenAI under the Apache 2.0 license. It uses a Mixture-of-Experts (MoE) architecture with 3.6B active parameters per forward pass, optimized for...

open-weightslow-cost
Input price
$0.02 / 1M tokens
Output price
$0.09 / 1M tokens
Context window
131K tokens
Max output
118K tokens
Output speed
109 tokens/s
Time to first token
1.19s
Inputs
text
Open weights
Yes
OpenRouter id
openai/gpt-oss-20b
Listed
Aug 5, 2025

Benchmarks

Intelligence Index
n/a
Engineering Index
n/a

Speed and benchmarks for “gpt-oss-20b (High)” from Artificial Analysis. Prices, context window and listing date are live from OpenRouter.

Compare with others