# Safety GPT OSS 20B

> Safety GPT OSS 20B is an open-weight language model from OpenAI, released on 29 Oct 2025. It accepts text and generates text, with a 131K-token context window and up to 66K output tokens per response. As of 4 Oct 2026, the lowest listed API price is $0.07 per million input tokens and $0.20 per million output tokens (Amazon Bedrock), across 4 providers we track.

Source page: https://www.aimodel.directory/models/gpt-oss-safeguard-20b
Last verified: 4 Oct 2026

## Specifications

- **Developer:** OpenAI
- **Release date:** 29 Oct 2025 (source: models.dev, https://github.com/sst/models.dev/blob/dev/providers/groq/models/openai/gpt-oss-safeguard-20b.toml)
- **Status:** Preview (source: models.dev, https://github.com/sst/models.dev/blob/dev/providers/groq/models/openai/gpt-oss-safeguard-20b.toml)
- **Weights:** Open weights (source: Hugging Face, https://huggingface.co/openai/gpt-oss-safeguard-20b)
- **License:** apache-2.0 (source: Hugging Face, https://huggingface.co/openai/gpt-oss-safeguard-20b)
- **Parameters:** 21.5B (source: Hugging Face, https://huggingface.co/openai/gpt-oss-safeguard-20b)
- **Context window:** 131K (131,072 tokens) (source: models.dev, https://github.com/sst/models.dev/blob/dev/providers/groq/models/openai/gpt-oss-safeguard-20b.toml)
- **Max output:** 66K tokens (source: models.dev, https://github.com/sst/models.dev/blob/dev/providers/groq/models/openai/gpt-oss-safeguard-20b.toml)
- **Input:** Text (source: models.dev, https://github.com/sst/models.dev/blob/dev/providers/groq/models/openai/gpt-oss-safeguard-20b.toml)
- **Output:** Text (source: models.dev, https://github.com/sst/models.dev/blob/dev/providers/groq/models/openai/gpt-oss-safeguard-20b.toml)
- **Reasoning mode:** Yes (source: models.dev, https://github.com/sst/models.dev/blob/dev/providers/groq/models/openai/gpt-oss-safeguard-20b.toml)
- **Tool calling:** Yes (source: models.dev, https://github.com/sst/models.dev/blob/dev/providers/groq/models/openai/gpt-oss-safeguard-20b.toml)
- **Structured output:** Yes (source: models.dev, https://github.com/sst/models.dev/blob/dev/providers/groq/models/openai/gpt-oss-safeguard-20b.toml)
- **Hugging Face:** openai/gpt-oss-safeguard-20b (source: Hugging Face, https://huggingface.co/openai/gpt-oss-safeguard-20b)

## API pricing (USD per 1M tokens)

| Provider | Input | Output | Cached input | Context |
|---|---|---|---|---|
| Amazon Bedrock | $0.07 | $0.20 | – | 128K |
| Vercel AI Gateway | $0.07 | $0.20 | – | 128K |
| Groq | $0.07 | $0.30 | – | 131K |
| OpenRouter | $0.07 | $0.30 | $0.04 | 131K |

## FAQ

### What is the context window of Safety GPT OSS 20B?

Safety GPT OSS 20B has a context window of 131,072 tokens (131K) and can generate up to 65,536 tokens in a single response, according to models.dev.

### How much does the Safety GPT OSS 20B API cost?

Safety GPT OSS 20B is listed from $0.07 per million input tokens and $0.20 per million output tokens (Amazon Bedrock). Prices last verified 4 Oct 2026.

### Is Safety GPT OSS 20B open source?

Safety GPT OSS 20B is an open-weight model: its weights are published on Hugging Face as openai/gpt-oss-safeguard-20b under the apache-2.0 license. Check the license terms before commercial use.

### When was Safety GPT OSS 20B released?

OpenAI released Safety GPT OSS 20B on 29 Oct 2025, according to models.dev.

### Which providers offer Safety GPT OSS 20B?

We track Safety GPT OSS 20B on 4 providers: Amazon Bedrock, Vercel AI Gateway, Groq, OpenRouter.

---
Data licensed CC BY 4.0. Cite as: AI Model Directory, "Safety GPT OSS 20B", https://www.aimodel.directory/models/gpt-oss-safeguard-20b