gpt-oss-safeguard-20b
gpt-oss-safeguard-20b is an open-weight language model from OpenAI with 21.5B parameters (mixture of experts). At 4-bit it needs about 14.4 GB of memory, so the smallest current device that fits it is the GeForce RTX 5060 Ti 16GB.
Overview
Includes room for an 8k context and runtime overhead. More context needs more memory.
Key information
- Lab
- OpenAI
- Released
- 18 Sep 2025
- Parameters
- 21.5B · MoE
- Type
- Language
- Context
- 131,072 tokens
- License
- Apache-2.0
- HF downloads, 30 days
- 24,864
- Repository
- openai/gpt-oss-safeguard-20b
More from OpenAI
View allWill it run on your hardware?
How we estimate| Estimate or device | Memory |
|---|---|
| Q4 | 14.4 GB |
| Q8 | 25.3 GB |
| BF16 | 47.5 GB |
| RTX 5060 Ti 16 GB | 16 GB usable |
| RTX 5090 32 GB | 32 GB usable |
| DGX Spark 128 GB | 108.8 GB usable |
| Mac Studio M5 Ultra 256 GB | 192 GB usable |
Single-device memory estimate. Pick a device to update the On your device column.
36 GB usable memory
| Format | Memory needed | On your device | Example hardware |
|---|---|---|---|
| 4-bit (Q4) | ≈ 14.4 GB | Fits 40% | GeForce RTX 5060 Ti 16GB16 GB memory |
| 8-bit (Q8) | ≈ 25.3 GB | Fits 70% | Radeon AI PRO R970032 GB memory |
| 16-bit (BF16) | ≈ 47.5 GB | Too big 132% | A100 80GB SXM80 GB memory |
Fits means the estimate stays within 90% of usable memory. Tight is above 90% and up to 100%. Too big is above 100%. These are memory checks, not speed or compatibility tests.
Mac allowances are 66% of total RAM up to 36 GB, and 75% above. AI boxes allow 85% because their memory is shared with the operating system.
Where to run it
On a Mac
Start with the MLX files, built for Apple Silicon.
lmstudio-community/gpt-oss-safeguard-20b-MLX-MXFP4Check your MacWith GGUF
Choose a file for llama.cpp or a compatible app.
lmstudio-community/gpt-oss-safeguard-20b-GGUFFind compatible hardwareThrough an API
Groq is the cheapest provider we found.
$0.075 / $0.30
Input / output per 1M tokens
₹7.07 / ₹28 at ₹94.29 per $1
Blended: $0.1312 (₹12) per 1M. Ranked by a 3:1 mix of input and output tokens.
mlx_lm.generate --model lmstudio-community/gpt-oss-safeguard-20b-MLX-MXFP4 \
--prompt "Explain how a transformer works."llama-cli -hf lmstudio-community/gpt-oss-safeguard-20b-GGUF \
-c 8192 -p "Explain how a transformer works."Example commands built from the repository ids above. They need a working install of MLX or llama.cpp. We have not run them; check each repository for available files.
| Precision | |||||
|---|---|---|---|---|---|
| GroqCheapest | $0.1312₹12 | $0.075₹7.07 | $0.30₹28 | 131,072 | Not listed |
INR at ₹94.29 per $1. Prices fetched 14 Nov 2026.
More about this model
Frequently asked questions
How much memory does gpt-oss-safeguard-20b need? At an 8k context, about 14.4 GB at 4-bit, 25.3 GB at 8-bit and 47.5 GB at 16-bit (BF16). A longer context needs more.
Can I run gpt-oss-safeguard-20b on a Mac? The smallest current Mac that fits the 4-bit estimate of 14.4 GB is the Mac mini M6 (32GB). An MLX build is available: lmstudio-community/gpt-oss-safeguard-20b-MLX-MXFP4.
Does the full context fit? gpt-oss-safeguard-20b supports up to 131,072 tokens of context. Our memory estimates use 8k, so long prompts need more memory than shown.
What does it cost through an API? The cheapest provider we found is Groq, at $0.08 per 1M input tokens and $0.30 per 1M output tokens (₹7.07 and ₹28 at 94.29 rupees per dollar). Ranked by a 3:1 mix of input and output tokens.
License in plain English
Apache-2.0. This license permits commercial use, modification and redistribution, with conditions such as keeping the required notices. This summary is not legal advice.
OMHQ RigBench
Not yet tested. We show no speed or quality score without a measurement.
Revision history and sources
Last updated . Sources:Hugging Face model card, MLX repository, GGUF repository, OpenRouter.