Skip to content

Nemotron 3.5 Lightning 30B A3B VRAM requirements

How much VRAM Nemotron 3.5 Lightning 30B A3B needs
4-bit weights: 17.8 GBWith 8K context: 18.7 GBWith 32K context: 18.8 GB8-bit, 8K context: 32.1 GB
Parameters
31.6B
Active per token
3B
Longest context
256K
Released
Aug 2026
License
OpenMDW 1.1
Strongest at
Agents, Docs

How much VRAM Nemotron 3.5 Lightning 30B A3B needs

Memory Nemotron 3.5 Lightning 30B A3B needs at each precision and context length
PrecisionWeightsWith 8KWith 32KWith 128K
16-bit58.8 GB59.7 GB59.8 GB60.4 GB
8-bit31.2 GB32.1 GB32.3 GB32.8 GB
6-bit24.1 GB25.0 GB25.1 GB25.7 GB
5-bit20.9 GB21.8 GB21.9 GB22.5 GB
4-bit17.8 GB18.7 GB18.8 GB19.4 GB
3-bit14.4 GB15.2 GB15.4 GB16.0 GB
2-bit11.0 GB11.9 GB12.0 GB12.6 GB
1.6-bit6.1 GB6.9 GB7.1 GB7.6 GB

One conversation, 16-bit context cache, runtime overhead included.

GPUs and Macs that run Nemotron 3.5 Lightning 30B A3B

The best way Nemotron 3.5 Lightning 30B A3B runs on common GPUs, Macs and AI boxes
MachineBest fitContextSpeedFirst replyReport
RTX 4060Too bigPDF: Nemotron 3.5 Lightning 30B A3B on RTX 4060 report
RTX 5060Too bigPDF: Nemotron 3.5 Lightning 30B A3B on RTX 5060 report
RTX 3070Too bigPDF: Nemotron 3.5 Lightning 30B A3B on RTX 3070 report
RTX 3080 10 GBToo bigPDF: Nemotron 3.5 Lightning 30B A3B on RTX 3080 10 GB report
RTX 3060 12 GBToo bigPDF: Nemotron 3.5 Lightning 30B A3B on RTX 3060 12 GB report
Arc B580Too bigPDF: Nemotron 3.5 Lightning 30B A3B on Arc B580 report
RTX 4070Too bigPDF: Nemotron 3.5 Lightning 30B A3B on RTX 4070 report
RTX 4070 SuperToo bigPDF: Nemotron 3.5 Lightning 30B A3B on RTX 4070 Super report
RTX 5070Too bigPDF: Nemotron 3.5 Lightning 30B A3B on RTX 5070 report
RTX 4060 Ti 16 GBToo bigPDF: Nemotron 3.5 Lightning 30B A3B on RTX 4060 Ti 16 GB report
RTX 5060 Ti 16 GBToo bigPDF: Nemotron 3.5 Lightning 30B A3B on RTX 5060 Ti 16 GB report
Arc A770 16 GBToo bigPDF: Nemotron 3.5 Lightning 30B A3B on Arc A770 16 GB report
Radeon RX 9070 XTToo bigPDF: Nemotron 3.5 Lightning 30B A3B on Radeon RX 9070 XT report
RTX 4070 Ti SuperToo bigPDF: Nemotron 3.5 Lightning 30B A3B on RTX 4070 Ti Super report
RTX 4080 SuperToo bigPDF: Nemotron 3.5 Lightning 30B A3B on RTX 4080 Super report
RTX 5070 TiToo bigPDF: Nemotron 3.5 Lightning 30B A3B on RTX 5070 Ti report
RTX 5080Too bigPDF: Nemotron 3.5 Lightning 30B A3B on RTX 5080 report
Radeon RX 7900 XTX5-bit176K118 tok/s0.8 sPDF: Nemotron 3.5 Lightning 30B A3B on Radeon RX 7900 XTX report
RTX 30905-bit176K155 tok/s0.5 sPDF: Nemotron 3.5 Lightning 30B A3B on RTX 3090 report
RTX 40905-bit176K161 tok/s0.2 sPDF: Nemotron 3.5 Lightning 30B A3B on RTX 4090 report
Radeon AI PRO R97006-bit256K89 tok/s0.5 sPDF: Nemotron 3.5 Lightning 30B A3B on Radeon AI PRO R9700 report
RTX 50906-bit256K195 tok/s0.2 sPDF: Nemotron 3.5 Lightning 30B A3B on RTX 5090 report
2× RTX 30908-bit256K124 tok/s0.3 sPDF: Nemotron 3.5 Lightning 30B A3B on 2× RTX 3090 report
2× RTX 40908-bit256K130 tok/s0.1 sPDF: Nemotron 3.5 Lightning 30B A3B on 2× RTX 4090 report
RTX PRO 5000 Blackwell 48 GB8-bit256K152 tok/s0.3 sPDF: Nemotron 3.5 Lightning 30B A3B on RTX PRO 5000 Blackwell 48 GB report
RTX 6000 Ada8-bit256K126 tok/s0.2 sPDF: Nemotron 3.5 Lightning 30B A3B on RTX 6000 Ada report
L40S8-bit256K118 tok/s0.3 sPDF: Nemotron 3.5 Lightning 30B A3B on L40S report
2× RTX 50908-bit256K175 tok/s0.1 sPDF: Nemotron 3.5 Lightning 30B A3B on 2× RTX 5090 report
4× RTX 30908-bit256K124 tok/s0.2 sPDF: Nemotron 3.5 Lightning 30B A3B on 4× RTX 3090 report
RTX PRO 6000 Blackwell8-bit256K175 tok/s0.1 sPDF: Nemotron 3.5 Lightning 30B A3B on RTX PRO 6000 Blackwell report
Mac mini M4 16 GBToo bigPDF: Nemotron 3.5 Lightning 30B A3B on Mac mini M4 16 GB report
Mac mini M4 24 GB3-bit32K44 tok/s12 sPDF: Nemotron 3.5 Lightning 30B A3B on Mac mini M4 24 GB report
Mac mini M4 32 GB4-bit256K38 tok/s12 sPDF: Nemotron 3.5 Lightning 30B A3B on Mac mini M4 32 GB report
Mac mini M4 Pro 24 GB3-bit32K71 tok/s5.9 sPDF: Nemotron 3.5 Lightning 30B A3B on Mac mini M4 Pro 24 GB report
Mac mini M4 Pro 48 GB8-bit256K46 tok/s5.9 sPDF: Nemotron 3.5 Lightning 30B A3B on Mac mini M4 Pro 48 GB report
Mac mini M4 Pro 64 GB8-bit256K46 tok/s5.9 sPDF: Nemotron 3.5 Lightning 30B A3B on Mac mini M4 Pro 64 GB report
M4 Max 128 GB8-bit256K66 tok/s2.9 sPDF: Nemotron 3.5 Lightning 30B A3B on M4 Max 128 GB report
M5 Max 128 GB8-bit256K70 tok/s0.9 sPDF: Nemotron 3.5 Lightning 30B A3B on M5 Max 128 GB report
M3 Ultra 96 GB8-bit256K75 tok/s1.7 sPDF: Nemotron 3.5 Lightning 30B A3B on M3 Ultra 96 GB report
M3 Ultra 256 GB8-bit256K75 tok/s1.7 sPDF: Nemotron 3.5 Lightning 30B A3B on M3 Ultra 256 GB report
M3 Ultra 512 GB8-bit256K75 tok/s1.7 sPDF: Nemotron 3.5 Lightning 30B A3B on M3 Ultra 512 GB report
M5 Ultra 512 GB8-bit256K87 tok/s0.5 sPDF: Nemotron 3.5 Lightning 30B A3B on M5 Ultra 512 GB report
Ryzen AI Max+ 395 64 GB8-bit256K39 tok/s1.7 sPDF: Nemotron 3.5 Lightning 30B A3B on Ryzen AI Max+ 395 64 GB report
Ryzen AI Max+ 395 128 GB8-bit256K39 tok/s1.7 sPDF: Nemotron 3.5 Lightning 30B A3B on Ryzen AI Max+ 395 128 GB report
RTX Spark8-bit256K54 tok/s0.6 sPDF: Nemotron 3.5 Lightning 30B A3B on RTX Spark report
DGX Spark8-bit256K50 tok/s0.7 sPDF: Nemotron 3.5 Lightning 30B A3B on DGX Spark report
Dual DGX Spark8-bit256K50 tok/s0.4 sPDF: Nemotron 3.5 Lightning 30B A3B on Dual DGX Spark report
DGX Station GB3008-bit256K265 tok/s0.1 sPDF: Nemotron 3.5 Lightning 30B A3B on DGX Station GB300 report
2× RTX PRO 6000 Blackwell8-bit256K175 tok/s0.1 sPDF: Nemotron 3.5 Lightning 30B A3B on 2× RTX PRO 6000 Blackwell report
4× RTX PRO 6000 Blackwell8-bit256K175 tok/s0.1 sPDF: Nemotron 3.5 Lightning 30B A3B on 4× RTX PRO 6000 Blackwell report
8× RTX PRO 6000 Blackwell8-bit256K175 tok/s0.1 sPDF: Nemotron 3.5 Lightning 30B A3B on 8× RTX PRO 6000 Blackwell report
A100 80 GB SXM8-bit256K186 tok/s0.4 sPDF: Nemotron 3.5 Lightning 30B A3B on A100 80 GB SXM report
H100 SXM8-bit256K222 tok/s0.1 sPDF: Nemotron 3.5 Lightning 30B A3B on H100 SXM report
4× H200 NVL8-bit256K245 tok/s0.1 sPDF: Nemotron 3.5 Lightning 30B A3B on 4× H200 NVL report
8× H200 NVL8-bit256K245 tok/s0.1 sPDF: Nemotron 3.5 Lightning 30B A3B on 8× H200 NVL report

Best fit is the highest quality, up to 8-bit, that leaves 16K of context. First reply is the wait for a 2K prompt. Speeds are rough, plus or minus 25%.

How to run it

Ollama
ollama run nemotron-3.5-lightning:30b-a3b
llama.cpp
llama-server -hf unsloth/NVIDIA-Nemotron-3.5-Lightning-30B-A3B-GGUF
vLLM
vllm serve nvidia/NVIDIA-Nemotron-3.5-Lightning-30B-A3B-BF16
MLX (Mac)
mlx_lm.chat --model mlx-community/NVIDIA-Nemotron-3.5-Lightning-30B-A3B-4bit

Questions

How much VRAM does Nemotron 3.5 Lightning 30B A3B need?

Nemotron 3.5 Lightning 30B A3B needs 17.8 GB for its weights at 4-bit, 18.7 GB with an 8K context and 18.8 GB with 32K. At 8-bit, 32.1 GB.

What is the minimum VRAM for Nemotron 3.5 Lightning 30B A3B?

About 15.2 GB at 3-bit with an 8K context, the lowest precision we recommend for a model this size. The smallest machine here that runs it without spilling is the Radeon RX 7900 XTX.

How big is Nemotron 3.5 Lightning 30B A3B?

31.6B parameters (3B active per token): 58.8 GB of weights at 16-bit, 17.8 GB at 4-bit.

Can Nemotron 3.5 Lightning 30B A3B run on an RTX 4090?

Yes. At 5-bit it holds 176K of context and writes about 161 tokens a second.

More models