Can I run this LLM?
Can a NVIDIA B200 (192GB) run GPT-OSS 20B (MoE)?
GPT-OSS · Apache 2.0 · estimated for a 192.0 GB machine at 4K context.
Yes — it runs great.
Runs greatBest fit: GPT-OSS 20B (MoE) at FP16 · 13.8 GB weights · 318–477 tok/s on a NVIDIA B200 (192GB).
Arena Text Overall · style-controlled
1,317FP16 untestedgpt-oss-20b · published configuration reference · FP16 not tested
Arena Text Overall (style-controlled), exact published entry gpt-oss-20b, 2026-10-08: reference rating 1,317.3. Published confidence bounds 1,310.9–1,323.8. 10,418 votes. Published gpt-oss-20b configuration reference. Reasoning mode is not disclosed by the Arena dataset. Tested precision, runtime and sampling settings are not disclosed in this dataset. Selected FP16 has not been measured by this Arena entry. No quantization quality adjustment is applied.
Snapshot published 2026-10-08 · snapshot imported 2026-10-11 UTC · validated last-good data · manual refresh
Arena leaderboard dataset · Arena / lmarena-ai, Arena Leaderboard Dataset (CC BY 4.0). Selected Text Overall style-controlled rows; field names adapted, ratings unchanged. · dataset CC BY 4.0 · revision e16f90232986
- Total memory at 4K context
- 14.7 GB
- Weights + KV cache + runtime overhead
- Estimated generation
- 318–477 tok/s
- Estimated first token
- ~0.1s
Every quantization on this machine
On small screens, each quant lists every metric below its name. On wider screens, scroll sideways for total memory, speed and fit. Keyboard: focus the table, then use the arrow keys. Quant names stay visible.
| Quant | Weights | Total memory | Speed | Fit |
|---|---|---|---|---|
| Q2_K | 11.5 GB | 12.4 GB | 325–488 tok/s | Runs great |
| Q3_K_M | 11.5 GB | 12.4 GB | 325–487 tok/s | Runs great |
| Q4_K_M | 11.6 GB | 12.5 GB | 325–487 tok/s | Runs great |
| Q5_K_M | 11.7 GB | 12.6 GB | 324–486 tok/s | Runs great |
| Q6_K | 12.0 GB | 12.9 GB | 323–485 tok/s | Runs great |
| Q8_0 | 12.1 GB | 13.0 GB | 323–485 tok/s | Runs great |
| FP16 | 13.8 GB | 14.7 GB | 318–477 tok/s | Runs great |
Ran GPT-OSS 20B (MoE) on a NVIDIA B200 (192GB)? Add your real numbers →