← AI Hashrate 中文

Can DGX Spark (GH200 480GB) run NVIDIA-Nemotron-3-Ultra-550B-A55B-NVFP4?

NVIDIA · 480 GB LPDDR5X · 500 GB/s bandwidth · NVIDIA · 550.0B (active 55.0B) · MoE · model ctx up to 256K

CC · WorkableComposite score: VRAM headroom × 0.5 + tok/s speed × 0.5. Q4 @ 8K context.
CUDA

Yes — the DGX Spark (GH200 480GB) (480 GB LPDDR5X) can run NVIDIA-Nemotron-3-Ultra-550B-A55B-NVFP4. At Q4 with the default 8K context it needs ≈ 304.05 GB VRAM (weights 302.5 GB + KV cache 0.55 GB + 1 GB runtime overhead), which fits within 480 GB. Decode at Q4/8K: ≈ 5.8 tok/s (estimated, batch 1). FP16 needs ≈ 1101.55 GB — does not fit on this card. Estimates come from memory-bandwidth math; rows tagged measured override estimates. Relative ranking is more reliable than absolute tok/s. Methodology. Buy DGX Spark (GH200 480GB) on Amazon →

Fit & speed by quant and context

QuantContextVRAM neededFits?tok/s (decode)
Q44K303.77 GBYes5.8 ± est.
Q48K default304.05 GBYes5.8 ± est.
Q432K305.7 GBYes5.8 ± est.
FP164K1101.28 GBNo0.9 ± est.
FP168K default1101.55 GBNo0.9 ± est.
FP1632K1103.2 GBNo0.9 ± est.

Fits = weights + KV(ctx) + 1 GB overhead ≤ 95% of VRAM. Measured anchors are context-agnostic; the fit verdict is recomputed per context. A missing tok/s means the model is far beyond this card (offload-only territory).

VRAM breakdown at 8K context

QuantWeightsKV cacheOverheadTotal neededDGX Spark (GH200 480GB) VRAM
Q4302.5 GB0.55 GB1 GB304.05 GB480 GB
FP161100.0 GB0.55 GB1 GB1101.55 GB480 GB

Other GPUs that run NVIDIA-Nemotron-3-Ultra-550B-A55B-NVFP4

No catalog GPU fully fits NVIDIA-Nemotron-3-Ultra-550B-A55B-NVFP4 at Q4/8K — see the model page for offload estimates. model page

All GPUs for NVIDIA-Nemotron-3-Ultra-550B-A55B-NVFP4 →

Other models for the DGX Spark (GH200 480GB)

All models on the DGX Spark (GH200 480GB) →