OpenAI · Released Oct 29, 2025
gpt-oss-safeguard-20b
gpt-oss-safeguard-20b is a safety reasoning model from OpenAI built upon gpt-oss-20b. This open-weight, 21B-parameter Mixture-of-Experts (MoE) model offers lower latency for safety tasks like content classification, LLM filtering, and…
- Released
- Oct 29 2025
- Context window
- 131K 66K out
- Input / 1M tokens
- $0.075 from $0.07
- Output / 1M tokens
- $0.30 from $0.20
Where to run it
9 providers compared
Cheapest: Vercel AI Gateway, Amazon Bedrock, Merge Gateway at $0.07 / $0.20 per 1M input / output tokens — 43% less than the priciest option.
| Provider | Input | Output | Cache read | Context | Max output | Uptime | vs. best |
|---|---|---|---|---|---|---|---|
| Vercel AI Gateway | $0.07 | $0.20 | — | 128K | 16K | — | Cheapest |
| Amazon Bedrock | $0.07 | $0.20 | — | 128K | 16K | — | Cheapest |
| Merge Gatewayno tool calling | $0.07 | $0.20 | Free | 4K | 4K | — | Cheapest |
| LowRouter | $0.08 | $0.24 | — | 131K | 131K | — | +17% |
| Groq | $0.075 | $0.30 | $0.038 | 131K | 66K | 99.6%24h | +28% |
| NanoGPTno tool calling | $0.075 | $0.30 | — | 128K | 16K | — | +28% |
| Kilo Gateway | $0.075 | $0.30 | $0.038 | 131K | 66K | — | +28% |
| Eden AI | $0.075 | $0.30 | $0.038 | 131K | 131K | — | +28% |
| AI/ML API | $0.10 | $0.41 | $0.052 | 131K | 66K | — | +76% |
Benchmarks
Benchmark scores
No benchmark scores yet — Epoch AI hasn't evaluated gpt-oss-safeguard-20b. Scores appear here automatically once it does.
Specs
Details
- Model ID
openai/gpt-oss-safeguard-20b- Developer
- OpenAI
- Release date
- Oct 29, 2025 ·
- Context window
- 131,072 tokens
- Max output
- 65,536 tokens
- API price
- $0.075 / $0.30 per 1M input / output tokens
- Providers
- 9
- Input types
- text
- Reasoning
- Yes
- Tool calling
- Yes
- Open weights
- Yes
- Free tier
- No
Release date and prices from OpenRouter's catalog; providers from OpenRouter and models.dev. Buying directly from a provider can cost a little differently. Speed, latency and uptime are measured on OpenRouter only; uptime is the share of OpenRouter requests each provider served over the last day. Benchmark scores are the best result per benchmark from Epoch AI, licensed CC BY 4.0. Rank counts every model tracked here with a score.