Only with CPU offload

Nemotron 3.5 Lightning 30B-A3B at Q4_K_M needs 18.4 GB but only 10.6 GB is addressable, so about 44% of the layers would stream from system RAM at roughly 60 GB/s. Expect around 15 tokens per second — usable for batch work, painful for chat.

56% on GPU 8K context Q4_K_M · 17.8 GB OpenMDW-1.1 Released 11 Aug 2026

Mamba-2 + MoE hybrid built for the execution layer of agents: only 6 attention blocks, so the KV cache is almost free. Weights, data and recipe all open.

Hardware requirements What hardware do I need to run Nemotron 3.5 Lightning 30B-A3B? →

At Q4_K_M with an 8K context: a 24 GB GPU (RTX 4090) or a 32 GB Mac (M4). Minimum, recommended, high-quality and long-context tiers, with all 114 devices checked at every quantisation.

The VRAM budget

weights 17.8 GB
Weights 17.8 GB KV cache @ 8K 0.05 GB Runtime overhead 0.6 GB Over budget 7.8 GB past 10.6 GB

Quantisation ladder

QuantWeightsTotal @ 8KMax contextTok/sQualityFit
Q8_0 31.3 GB 31.9 GB — ~6.1 −0.1% ppl 21.3 GB over
Q6_K 24.1 GB 24.8 GB — ~8.9 −0.4% ppl 14.2 GB over
Q5_K_M 20.9 GB 21.5 GB — ~11 −0.8% ppl 10.9 GB over
Q4_K_M 17.8 GB 18.4 GB — ~15 −1.9% ppl 7.8 GB over
Q3_K_M 14.4 GB 15.0 GB — ~24 −5.4% ppl 4.4 GB over
Q2_K 12.3 GB 13.0 GB — ~37 −15% ppl 2.4 GB over

Quality is the published perplexity delta against f16 weights. Max context assumes an f16 KV cache; q8_0 roughly doubles it. Only 6 of its 52 blocks keep a per-token KV cache; the rest are linear-attention, Mamba or convolution blocks with a fixed-size state.

How to run it

terminal
$ llama-server \
    -hf nvidia/NVIDIA-Nemotron-3.5-Lightning-30B-A3B-BF16:Q4_K_M \
    -c 8192 -ngl 29

The engine underneath most of the others. Every knob is exposed. More on llama.cpp.

01Download is 17.8 GB. Keep it on an SSD — a first load off a spinning disk takes minutes.
02Close anything else holding VRAM. A browser with hardware acceleration can sit on 1–2 GB.
03If it falls back to CPU silently, drop the context first, then step down a quantisation.
See all models for this rig Compare with another model

Related models on a GeForce RTX 4080 Laptop

ModelTotalTok/s est.
Nemotron 3 Super 120B-A12B
124B MoE · same family
Q4_K_M · 70.4 GB — Won't fit
Gemma 4 31B
31.3B · similar size
Q4_K_M · 20.2 GB ~3.4 CPU offload
GLM-4.7-Flash 30B-A3B
31.2B MoE · similar size
Q4_K_M · 18.6 GB ~16 CPU offload
Olmo 3.1 32B Instruct
32.2B · similar size
Q4_K_M · 20.0 GB ~3.4 CPU offload
Qwen3 30B-A3B
30.5B MoE · similar size
Q4_K_M · 18.5 GB ~14 CPU offload
Qwen3 Coder 30B-A3B
30.5B MoE · similar size
Q4_K_M · 18.5 GB ~14 CPU offload
Qwen3 32B
32.8B · similar size
Q4_K_M · 21.0 GB ~3.1 CPU offload

Every model against the GeForce RTX 4080 Laptop →

Nemotron 3.5 Lightning 30B-A3B on other hardware

DeviceTotalTok/s est.
GeForce RTX 4060 Laptop
7.0 GB usable · 256 GB/s
18.4 GB ~11 CPU offload
GeForce RTX 3060 12 GB
10.6 GB usable · 360 GB/s
18.4 GB ~15 CPU offload
M4 · 16 GB
10.7 GB usable · 120 GB/s
18.4 GB — Won't fit
GeForce RTX 4090
22.4 GB usable · 1008 GB/s
18.4 GB 130 Runs great
Radeon RX 7900 XTX
22.4 GB usable · 960 GB/s
18.4 GB 124 Runs great
M4 · 32 GB
24.0 GB usable · 120 GB/s
18.4 GB 17 Runs great
M3 Max · 64 GB
48.0 GB usable · 400 GB/s
18.4 GB 58 Runs great

All 114 devices, by tier →