gpt-oss-20b
Compact open-weight reasoning model that runs on consumer hardware.
gpt-oss-20b is an open-weight 21B-parameter model released under Apache 2.0. It uses a mixture-of-experts architecture with 3.6B active parameters, optimized for low-latency and local deployment.
- Provider
- OpenAI
- Type
- Language model
- Released
- Aug 5, 2025
- Context window
- 131K tokens
- Max output
- 33K tokens
- Price
- $0.018 input / $0.090 output per 1M tokens
- Input
- text
- Output
- text
- Open weights
- Yes
- License
- Apache 2.0
Best for
- Local & on-device
- High volume / low cost
- Real-time / low latency
- Extraction & classification
Strengths
- Runs on a laptop GPU
- Apache 2.0
- Near-zero hosted cost
More from OpenAI
- GPT-6 Astra — OpenAI. OpenAI’s flagship for deep research, science and end-to-end engineering.
- GPT-6.1 Sol — OpenAI. Mid-tier GPT-6 model with excellent cost per task for agentic coding.
- GPT-6 Luna — OpenAI. Fast, cost-efficient GPT-6 tier for chat, classification and light agents.
- GPT-5.6 Terra — OpenAI. Balanced GPT-5.6 model for everyday coding and reasoning.
- gpt-oss-120b — OpenAI. OpenAI’s Apache-2.0 open-weight reasoning model; runs on a single 80GB GPU.
- GPT-5.4 Image 2 — OpenAI. OpenAI’s native image generation and editing model.
- text-embedding-3-large — OpenAI. OpenAI’s most capable embedding; the default in countless RAG stacks.
- text-embedding-3-small — OpenAI. Cheap, fast OpenAI embedding for high-volume indexing.
Tools that use it
- Ollama — Ollama. Run open models locally with one command.