GPT-OSS 20B in MXFP4 with an 8k context takes 12.6 GB of GPU memory. The cheapest option right now is 2× RTX 3060 12 GB, from $0.174/h.
The weight file and the total GPU memory with the context cache. One GPU, one request.
| Quantisation | File | 8K context | 32K context | 128K context | Fits on |
|---|---|---|---|---|---|
| MXFP4native | 11.3 | 12.6 | 13.2 | 15.4 | RTX A4000 16 GB |
| BF16 | ≈38.9 | 40.3 | 40.8 | 43.1 | RTX A6000 48 GB |
How much memory the cache takes per request in FP16. A Q8 cache is about half. 24 KB per token + 128 × 24 KB.
GPUs from the catalog the model fits on, by price per hour.
The price is the cheapest GPU of the model × the GPU count. A 2–8 GPU server can cost a little more per GPU.
The “GPT-OSS 20B Uncensored” template starts the model in a ready environment: open the link once the server has loaded.
In the Ollama template the model is one command away:
ollama run gpt-oss:20bFor an API under load there is the vLLM template with an OpenAI-compatible API.
File sizes checked against Hugging Face on 2026-10-19. 1 GB here is 1024³ bytes, like GPU memory. Hugging Face lists decimal gigabytes, so its numbers are 7% larger.