Can I run this LLM?
Can a NVIDIA DGX Spark (GB10, 128GB) run Muse Glimmer 30B?
Muse · Apache 2.0 · estimated for a 128.0 GB machine at 4K context.
Yes — it runs great.
Runs greatBest fit: Muse Glimmer 30B at Q3_K_M · 14.0 GB weights · 12–18 tok/s on a NVIDIA DGX Spark (GB10, 128GB).
Arena Text Overall · style-controlled
UnratedAmbiguous matchAmbiguous Arena identity; unrated
Arena identity is ambiguous. The Arena identifier omits the 30B size. An independent Arena alias-to-release binding is not yet verified. No rating is assigned or used for ranking. Local quantization quality is unmeasured.
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
- 14.9 GB
- Weights + KV cache + runtime overhead
- Estimated generation
- 12–18 tok/s
- Estimated first token
- ~6.9s
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 | 11.0 GB | 12.0 GB | 15–23 tok/s | Runs great |
| Q3_K_M | 14.0 GB | 14.9 GB | 12–18 tok/s | Runs great |
| Q4_K_M | 17.3 GB | 18.2 GB | 10–15 tok/s | Tight / slow |
| Q5_K_M | 20.1 GB | 21.0 GB | 9–13 tok/s | Tight / slow |
| Q6_K | 23.4 GB | 24.3 GB | 7–11 tok/s | Tight / slow |
| Q8_0 | 29.6 GB | 30.5 GB | 6–9 tok/s | Tight / slow |
| FP16 | 55.7 GB | 56.7 GB | 3–5 tok/s | Tight / slow |
Ran Muse Glimmer 30B on a NVIDIA DGX Spark (GB10, 128GB)? Add your real numbers →