
OpenAI released gpt-oss-20b (free) on August 4, 2025. The registry currently records it as open weights.
Model publication
OpenAI
gpt-oss-20b (free)Open the source for gpt-oss-20b (free)

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 lower-latency inference and deployability on consumer or single-GPU hardware. The model is trained in OpenAI’s Harmony response format and supports reasoning level configuration, fine-tuning, and agentic capabilities including function calling, tool use, and structured outputs.
- Licence
- apache-2.0
- Architecture
- mixture_of_experts
- Parameters
- 21.5B
- Context
- 131.1K tokens
Benchmarks
No benchmark observation recorded for this model.