Can I Run GPT-OSS 20B on NVIDIA A100 80GB (PCIe)?
Written by Jakub Rusinowski · Last updated August 15, 2026
Yes. GPT-OSS 20B needs 12.3 GB at 8K context (11.1 GB weights + 1.2 GB KV cache/overhead) at MXFP4 and the NVIDIA A100 80GB (PCIe) has 80 GB, leaving about 67.7 GB of headroom, at roughly 250 tok/s (estimated).
See what else this hardware can run →
The numbers
| Model | GPT-OSS 20B |
| Parameters | 20B |
| Quantization | MXFP4 |
| VRAM needed (8K context) | 12.3 GB |
| NVIDIA A100 80GB (PCIe) VRAM | 80 GB |
| Verdict | Runs |
| Estimated speed | ~250 tok/s |
GPT-OSS 20B on the NVIDIA A100 80GB (PCIe): VRAM by quantization
| Quant | VRAM needed | Fits 80 GB? | Max context | Whole PC |
|---|---|---|---|---|
| F16 | 43 GB | ✓ Yes | 128K | — |
| Q8_0 | 23.4 GB | ✓ Yes | 128K | — |
| Q6_K | 18.3 GB | ✓ Yes | 128K | — |
| Q5_K_M | 16 GB | ✓ Yes | 128K | — |
| Q4_K_M | 13.8 GB | ✓ Yes | 128K | — |
| Q3_K_M | 10.1 GB | ✓ Yes | 128K | — |
| Q2_K | 8.1 GB | ✓ Yes | 128K | — |
Assumes an 8K-token context with an f16 KV cache. A longer window needs more; a quantized KV cache needs less. “Max context” is the largest window that still fits in 80 GB. Figures are estimates from parameter count, quantization and memory bandwidth — the analyzer lets you tune KV-cache quant and context.
Context costs VRAM too. At MXFP4 on the NVIDIA A100 80GB (PCIe): 8K 12.3 GB · 32K 13.5 GB · 128K 18.3 GB. Every window this model publishes still fits 80 GB.
What to do next
Nothing to change — run it at MXFP4. If you want more capability, check whether the next size up also fits.
No compatible GPU? GPT-OSS 20B on 16 GB of system RAM, CPU only: runs.
Running GPT-OSS 20B on this card:
At 2 hrs/day, buying (~$15,000) beats renting at $0.34/hr after about 61.3 years.
Affiliate links — we may earn a commission if you sign up, at no extra cost to you.
Cloud rates verified 2026-07 — estimates, and marketplace prices vary. Buying price is GPU MSRP only, not a full PC.
Other GPT-OSS sizes on the NVIDIA A100 80GB (PCIe)
- GPT-OSS 120B — runs (71.9 GB)
GPT-OSS 20B on nearby hardware
- Apple M2 (24 GB) — runs
- Apple M2 Max (96 GB) — runs
All GPT-OSS sizes on the NVIDIA A100 80GB (PCIe) | GPT-OSS 20B model page | NVIDIA A100 80GB (PCIe) GPU page | VRAM calculator | Check your hardware