← AI Hashrate 中文

Can DGX Spark (GH200 480GB) run MiniMax-M3?

NVIDIA · 480 GB LPDDR5X · 500 GB/s bandwidth · MiniMax · 428.0B (active 23.0B) · MoE · model ctx up to 1024K

✅ Yes — DGX Spark (GH200 480GB) (480 GB) can run MiniMax-M3 at Q4 (8K)
VRAM: 241.0 GB needed of 480 GB
Speed: ~13.8 tok/s (estimated)
BB · GoodComposite score: VRAM headroom × 0.5 + tok/s speed × 0.5. Q4 @ 8K context.
CUDA
Buy DGX Spark (GH200 480GB) on Amazon →
Estimates from memory-bandwidth formula; rows marked "measured" override. Methodology.

Fit & speed by quant and context

QuantContextVRAM neededFits?tok/s (decode)
Q44K238.7 GBYes13.8 ± est.
Q48K default241.0 GBYes13.8 ± est.
Q432K254.8 GBYes13.8 ± est.
FP164K859.3 GBNo3.4 ± est.
FP168K default861.6 GBNo3.4 ± est.
FP1632K875.4 GBNo3.3 ± 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
Q4235.4 GB4.6 GB1 GB241.0 GB480 GB
FP16856.0 GB4.6 GB1 GB861.6 GB480 GB

Other GPUs that run MiniMax-M3

No catalog GPU fully fits MiniMax-M3 at Q4/8K — see the model page for offload estimates. model page

All GPUs for MiniMax-M3 →

Other models for the DGX Spark (GH200 480GB)

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