Hardware requirements

What hardware do I need to run Nemotron 3 Super 120B-A12B?

124B MoE · 12B active 256K context NVIDIA Open Model

At Q4_K_M the weights are 69.7 GB; with an 8K KV cache and runtime overhead that is 70.4 GB, so it wants a 141 GB card to run with headroom — in practice 141 GB VRAM: H200 SXM. Every figure below is computed from the model’s published shape against each device’s usable memory; tokens per second are estimates. The method.

Minimum practical 64 GB VRAM

Instinct MI210 — Q3_K_M, 57.1 GB of 62.4 GB, ~70 tok/s, up to 256K context.

the smallest card that holds it at all, at Q3_K_M — the lowest quantisation worth running
Recommended 141 GB VRAM

H200 SXM — Q4_K_M, 70.4 GB of 139.4 GB, ~166 tok/s, up to 256K context.

runs the recommended quantisation with headroom at 8K context
High quality 192 GB VRAM

Instinct MI300X — Q8_0, 123.4 GB of 190.4 GB, ~104 tok/s, up to 256K context.

near-lossless Q8_0 with headroom
Long context 141 GB VRAM

H200 SXM — Q4_K_M, 71.3 GB of 139.4 GB, ~166 tok/s at 128K.

128K tokens (or the model’s maximum) at the recommended quantisation
Apple Silicon 128 GB unified

M1 Ultra · 128 GB — Q4_K_M, 70.4 GB of 96.0 GB, ~31 tok/s, up to 256K context.

Also: M4 Max · 128 GB, M3 Max · 128 GB, Ryzen AI Max+ 395 · 128 GB
smallest Mac that runs the recommended quantisation with headroom
CPU only 256 GB RAM

CPU only · DDR5 8-channel server — Q8_0, 123.4 GB of 204.8 GB, ~4.8 tok/s, up to 256K context.

no GPU — weights stream from system RAM

How much VRAM at each quantisation

QuantWeights+KV 8K+KV 32K+KV 128KTotal @ 8KNeedsQuality
Q8_0122.7 GB0.06 GB0.25 GB1.0 GB 123.4 GB192 GB card−0.1% ppl
Q6_K94.7 GB0.06 GB0.25 GB1.0 GB 95.4 GB141 GB card−0.4% ppl
Q5_K_M81.8 GB0.06 GB0.25 GB1.0 GB 82.5 GB141 GB card−0.8% ppl
Q4_K_M69.7 GB0.06 GB0.25 GB1.0 GB 70.4 GB141 GB card−1.9% ppl
Q3_K_M56.4 GB0.06 GB0.25 GB1.0 GB 57.1 GB80 GB card−5.4% ppl
Q2_K48.4 GB0.06 GB0.25 GB1.0 GB 49.0 GB64 GB card−15% ppl

Totals add 0.6 GB runtime overhead and an f16 KV cache; "needs" is the smallest discrete card class whose usable memory (after display and driver reserve) leaves 15% headroom. A q8_0 cache roughly halves the KV columns.

Every device, checked

System RAM for offload:
163264128256

NVIDIA · GeForce

DeviceMemoryBest quant with headroomFits at allMax context @ Q4_K_MTok/s est.128K
GeForce RTX 5090 32 GB — offload — ~5.2 no Check
GeForce RTX 5080 16 GB — no — — no Check
GeForce RTX 5070 Ti 16 GB — no — — no Check
GeForce RTX 5070 12 GB — no — — no Check
GeForce RTX 5060 Ti 16 GB 16 GB — no — — no Check
GeForce RTX 5060 8 GB — no — — no Check
GeForce RTX 4090 24 GB — no — — no Check
GeForce RTX 4080 Super 16 GB — no — — no Check
GeForce RTX 4080 16 GB — no — — no Check
GeForce RTX 4070 Ti Super 16 GB — no — — no Check
GeForce RTX 4070 Ti 12 GB — no — — no Check
GeForce RTX 4070 Super 12 GB — no — — no Check
GeForce RTX 4070 12 GB — no — — no Check
GeForce RTX 4060 Ti 16 GB 16 GB — no — — no Check
GeForce RTX 4060 Ti 8 GB — no — — no Check
GeForce RTX 4060 8 GB — no — — no Check
GeForce RTX 3090 Ti 24 GB — no — — no Check
GeForce RTX 3090 24 GB — no — — no Check
GeForce RTX 3080 Ti 12 GB — no — — no Check
GeForce RTX 3080 12 GB 12 GB — no — — no Check
GeForce RTX 3080 10 GB — no — — no Check
GeForce RTX 3070 Ti 8 GB — no — — no Check
GeForce RTX 3070 8 GB — no — — no Check
GeForce RTX 3060 Ti 8 GB — no — — no Check
GeForce RTX 3060 12 GB 12 GB — no — — no Check
GeForce RTX 2080 Ti 11 GB — no — — no Check
GeForce RTX 2060 12 GB 12 GB — no — — no Check
GeForce GTX 1080 Ti 11 GB — no — — no Check

NVIDIA · GeForce laptop

DeviceMemoryBest quant with headroomFits at allMax context @ Q4_K_MTok/s est.128K
GeForce RTX 5090 Laptop 24 GB — no — — no Check
GeForce RTX 4090 Laptop 16 GB — no — — no Check
GeForce RTX 4080 Laptop 12 GB — no — — no Check
GeForce RTX 4070 Laptop 8 GB — no — — no Check

NVIDIA · Workstation

DeviceMemoryBest quant with headroomFits at allMax context @ Q4_K_MTok/s est.128K
RTX 6000 Ada 48 GB — offload — ~11 no Check
RTX 5000 Ada 32 GB — offload — ~4.8 no Check
RTX 4000 Ada 20 GB — no — — no Check
RTX 2000 Ada 16 GB — no — — no Check
RTX A6000 48 GB — offload — ~10 no Check
RTX A5000 24 GB — no — — no Check
RTX A4000 16 GB — no — — no Check

NVIDIA · Data centre

DeviceMemoryBest quant with headroomFits at allMax context @ Q4_K_MTok/s est.128K
L40S 48 GB — offload — ~10 no Check
L4 24 GB — no — — no Check
A10 24 GB — no — — no Check
A100 40 GB 40 GB — offload — ~7.2 no Check
A100 80 GB 80 GB Q3_K_M Q4_K_M 256K 70 tight Check
H100 PCIe 80 GB Q3_K_M Q4_K_M 256K 69 tight Check
H100 SXM 80 GB Q3_K_M Q4_K_M 256K 116 tight Check
H200 SXM 141 GB Q6_K Q8_0 256K 166 yes Check
Tesla P40 24 GB — no — — no Check

AMD · Radeon

DeviceMemoryBest quant with headroomFits at allMax context @ Q4_K_MTok/s est.128K
Radeon RX 9070 XT 16 GB — no — — no Check
Radeon RX 9070 16 GB — no — — no Check
Radeon RX 7900 XTX 24 GB — no — — no Check
Radeon RX 7900 XT 20 GB — no — — no Check
Radeon RX 7900 GRE 16 GB — no — — no Check
Radeon RX 7800 XT 16 GB — no — — no Check
Radeon RX 7700 XT 12 GB — no — — no Check
Radeon RX 7600 XT 16 GB — no — — no Check
Radeon RX 6900 XT 16 GB — no — — no Check
Radeon RX 6800 XT 16 GB — no — — no Check
Radeon RX 6700 XT 12 GB — no — — no Check

AMD · Radeon Pro

DeviceMemoryBest quant with headroomFits at allMax context @ Q4_K_MTok/s est.128K
Radeon Pro W7900 48 GB — offload — ~10 no Check
Radeon Pro W7800 32 GB — offload — ~4.8 no Check

AMD · Instinct

DeviceMemoryBest quant with headroomFits at allMax context @ Q4_K_MTok/s est.128K
Instinct MI210 64 GB — Q3_K_M — — no Check
Instinct MI300X 192 GB Q8_0 Q8_0 256K 183 yes Check

AMD · Ryzen AI Max

DeviceMemoryBest quant with headroomFits at allMax context @ Q4_K_MTok/s est.128K
Ryzen AI Max+ 395 · 128 GB 128 GB Q4_K_M Q6_K 256K 9.9 yes Check
Ryzen AI Max+ 395 · 64 GB 64 GB — no — — no Check

Apple · Apple Silicon

DeviceMemoryBest quant with headroomFits at allMax context @ Q4_K_MTok/s est.128K
M1 · 16 GB 16 GB — no — — no Check
M1 Pro · 32 GB 32 GB — no — — no Check
M1 Max · 64 GB 64 GB — no — — no Check
M1 Ultra · 128 GB 128 GB Q4_K_M Q6_K 256K 31 yes Check
M2 · 24 GB 24 GB — no — — no Check
M2 Pro · 32 GB 32 GB — no — — no Check
M2 Max · 96 GB 96 GB Q3_K_M Q4_K_M 214K 15 tight Check
M2 Ultra · 192 GB 192 GB Q6_K Q8_0 256K 31 yes Check
M3 · 24 GB 24 GB — no — — no Check
M3 Pro · 36 GB 36 GB — no — — no Check
M3 Max · 64 GB 64 GB — no — — no Check
M3 Max · 128 GB 128 GB Q4_K_M Q6_K 256K 15 yes Check
M3 Ultra · 256 GB 256 GB Q8_0 Q8_0 256K 31 yes Check
M3 Ultra · 512 GB 512 GB Q8_0 Q8_0 256K 31 yes Check
M4 · 32 GB 32 GB — no — — no Check
M4 Pro · 48 GB 48 GB — no — — no Check
M4 Pro · 64 GB 64 GB — no — — no Check
M4 Max · 64 GB 64 GB — no — — no Check
M4 Max · 128 GB 128 GB Q4_K_M Q6_K 256K 21 yes Check

Intel · Arc

DeviceMemoryBest quant with headroomFits at allMax context @ Q4_K_MTok/s est.128K
Arc B580 12 GB — no — — no Check
Arc B570 10 GB — no — — no Check
Arc A770 16 GB 16 GB — no — — no Check
Arc A750 8 GB — no — — no Check

CPU · CPU + system RAM

DeviceMemoryBest quant with headroomFits at allMax context @ Q4_K_MTok/s est.128K
CPU only · DDR4 dual-channel 32 GB — no — — no Check
CPU only · DDR5 dual-channel 64 GB — no — — no Check
CPU only · DDR5 8-channel server 256 GB Q8_0 Q8_0 256K 8.5 yes Check