Can I run this LLM?
Can a Mac Studio M3 Ultra (96GB) run Nemotron 3.5 Lightning 30B-A3B (MoE)?
Nemotron · openmdw-1.1 · estimated for a 96.0 GB machine at 4K context.
Yes — it runs great.
Runs greatBest fit: Nemotron 3.5 Lightning 30B-A3B (MoE) at Q8_0 · 35.0 GB weights · 86–129 tok/s on a Mac Studio M3 Ultra (96GB).
Arena Text Overall · style-controlled
UnratedAmbiguous matchAmbiguous Arena identity; unrated
Arena identity is ambiguous. The published NVFP4 configuration differs from generic local GGUF quantizations. 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
- 35.8 GB
- Weights + KV cache + runtime overhead
- Estimated generation
- 86–129 tok/s
- Estimated first token
- ~1.3s
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 | 18.9 GB | 19.7 GB | 111–167 tok/s | Runs great |
| Q3_K_M | 19.8 GB | 20.6 GB | 109–164 tok/s | Runs great |
| Q4_K_M | 25.5 GB | 26.3 GB | 99–149 tok/s | Runs great |
| Q5_K_M | 27.0 GB | 27.8 GB | 97–145 tok/s | Runs great |
| Q6_K | 34.3 GB | 35.1 GB | 87–130 tok/s | Runs great |
| Q8_0 | 35.0 GB | 35.8 GB | 86–129 tok/s | Runs great |
| FP16 | 65.9 GB | 66.7 GB | 60–90 tok/s | Tight / slow |
Ran Nemotron 3.5 Lightning 30B-A3B (MoE) on a Mac Studio M3 Ultra (96GB)? Add your real numbers →