OpenAI: gpt-oss-safeguard-20b
openai/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 trust & safety labeling.
Learn more about this model in OpenAI's gpt-oss-safeguard user guideOpens in new tab.
Providers
Different companies host the same model. OpenRouter routes your request to one of them based on the routing mode you pick: Balanced (price + speed), Nitro (fastest), or Floor (cheapest).
| $0.070 | $0.20 | -- | 0.36s | 212 tok/s | ||
| $0.075 | $0.30 | $0.0375 | 0.17s | 603 tok/s |
- Input /M
- $0.07
- Output /M
- $0.20
- Cache read /M
- --
- Latency
- 0.36s
- Throughput
- 212 tok/s
- Uptime
- 99.30%
- Input /M
- $0.075
- Output /M
- $0.30
- Cache read /M
- $0.0375
- Latency
- 0.17s
- Throughput
- 603 tok/s
- Uptime
- 81.30%
Pricing
The average price customers actually pay for this model, next to the prices providers post. Caching and discounts mean the price actually paid is often well below the listed one.
Weighted average
gpt-oss-safeguard-20b — Price history
| $0.0697 | $0.199 | 0.0% | 28.4% | ||
| $0.0672 | $0.299 | 20.4% | 71.6% |
- Effective in $0.0697 / Effective out $0.199Token share28.4%
- Effective in $0.0672 / Effective out $0.299Token share71.6%
Performance
Throughput is how fast the model writes (tokens per second — higher is better). Latency is total round-trip time (lower is better). TTFT is time-to-first-token — how long before you see anything appear (lower is better).
P50, best across providers
P50, best provider
Uptime
Uptime is the percentage of the past 3 days that at least one provider was responding to requests. Availability is the percentage of time that inference was successfully served. OpenRouter continuously monitors and uses the next-best provider when one returns an error.
Availability over the last 3 days
Last 72 hoursAvailability over the last 24 hours
When an error occurs in an upstream provider, we can recover by routing to another healthy provider, if your request filters allow it. You can access per-provider uptime data programmatically through the Endpoints API. Learn more about our load balancing and provider routing options.
Apps
Public apps that send the most traffic to this model. Good signal for what real production workloads look like — and a hint at which use cases this model is best suited for.
Activity
Token volume and request traffic to this model over time.
API
Drop-in code to call this model. OpenRouter's API is OpenAI-compatible — most SDKs work by just swapping the base URL. The only thing that changes between models is the model slug below.
Not currently available through EU in-region routing. In-region routing docs
Frequently asked questions
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 trust & safety labeling. Learn more about this model in OpenAI's gpt-oss-safeguard user guide.
gpt-oss-safeguard-20b costs $0.07/M input tokens and $0.20/M output tokens.
gpt-oss-safeguard-20b has a 131,072 token context window.
Yes. gpt-oss-safeguard-20b accepts tools and tool_choice for function calling. It also supports structured outputs via a JSON schema in response_format.
gpt-oss-safeguard-20b is served by 2 providers on OpenRouter: Amazon Bedrock and Groq. Requests are routed to the best available provider, with automatic failover to the others, and you can pin or exclude providers with provider routing.
gpt-oss-safeguard-20b was released on October 29, 2025.