Can I run this LLM?
Can a NVIDIA GB300 Grace Blackwell (288GB HBM + 480GB) run GPT-OSS 120B (MoE)?
GPT-OSS · Apache 2.0 · estimated for a 288.0 GB machine at 4K context.
Yes — it runs great.
Best fit: GPT-OSS 120B (MoE) at FP16 · 65.4 GB · 307–460 tok/s on a NVIDIA GB300 Grace Blackwell (288GB HBM + 480GB).
Every quantization on this machine
| Quant | Size | Speed | Fit |
|---|---|---|---|
| Q2_K | 62.6 GB | 309–463 tok/s | Runs great |
| Q3_K_M | 62.6 GB | 309–463 tok/s | Runs great |
| Q4_K_M | 62.8 GB | 309–463 tok/s | Runs great |
| Q5_K_M | 62.9 GB | 309–463 tok/s | Runs great |
| Q6_K | 63.3 GB | 308–462 tok/s | Runs great |
| Q8_0 | 63.4 GB | 308–462 tok/s | Runs great |
| FP16 | 65.4 GB | 307–460 tok/s | Runs great |
Open the full calculator →Everything a NVIDIA GB300 Grace Blackwell (288GB HBM + 480GB) runsGPT-OSS 120B (MoE) on other machinesWhat to buy to run it →
Ran GPT-OSS 120B (MoE) on a NVIDIA GB300 Grace Blackwell (288GB HBM + 480GB)? Add your real numbers →