Nemotron 3 Nano 30B-A3B VRAM requirements

Can I run Nemotron 3 Nano 30B-A3B on…

• RTX 4090: Yes, Q4_K_M
• RTX 3090: Yes, Q4_K_M
• RTX 5090: Yes, up to Q6_K
• RTX 3060 12GB: --n-cpu-moe 30
• RTX 4060 Ti 16GB: Only Q2_K
• RTX 5060 Ti 16GB: Only Q2_K
• RTX 4070 12GB: --n-cpu-moe 30
• RTX 5080 16GB: Only Q2_K
• RTX 4080 Super 16GB: Only Q2_K
• RX 7900 XTX: Yes, Q4_K_M

Running Nemotron 3 Nano 30B-A3B with experts in system RAM

llama.cpp's --n-cpu-moe N keeps the experts of the first N layers in RAM. For the measured Q4_K_M GGUF with 32K tokens of context, the smallest N that fits each card, and the speed with 89.6 GB/s of system RAM. Plan the offload for another GPU. The VRAM here can sit below the 20.3 GB of the worked example because it counts the file's measured bytes, keeps the input embedding (token_embd) in system RAM as llama.cpp does and adds a flat 1 GB of buffers, where the estimate above adds 0.5 GB plus 10% to the preset's weights.

In system RAM counts the offloaded experts plus the input embedding (token_embd), which stays in RAM even with --n-cpu-moe 0.

How much VRAM does Nemotron 3 Nano 30B-A3B need?

Nemotron 3 Nano 30B-A3B needs 20.1 GB at Q4_K_M, 34.9 GB at Q8_0 and 65.3 GB at FP16 with 8,192 tokens of context and one request. Each total below is the weights plus the FP16 KV cache and the runtime overhead (0.5 GB plus 10%), and each row opens the calculator with that setting. What the GGUF names mean.

Fine-tuning Nemotron 3 Nano 30B-A3B? Nemotron 3 Nano 30B-A3B VRAM for LoRA, QLoRA and full training.

Nemotron 3 Nano 30B-A3B GGUF files on Hugging Face

The 8 GGUF files llama.cpp picks for Nemotron 3 Nano 30B-A3B with -hf <repo>:<QUANT>, with the bytes the Hugging Face file list reported on 2026-09-29. Each total is that file plus the FP16 KV cache and 0.5 GB + 10%, one request. "vs estimate" compares the file with the bits-per-weight size the table above uses.

GB here is GiB (1024³ bytes), as everywhere on this site; Hugging Face shows the Q4_K_M file as 22.4 GB, in 10⁹ bytes.

Which Nemotron 3 Nano 30B-A3B GGUF fits a 16, 24, 32, 48 or 80 GB GPU or a Mac

From the file sizes above: the largest file that fits each machine with an 8K context and 0.5 GB left free, the longest context it then has room for, and the longest context for Q4_K_M (20.9 GB). One request, FP16 KV cache. Where nothing fits, the --n-cpu-moe table above keeps experts in system RAM.

Nemotron 3 Nano 30B-A3B at 8K, 32K, 128K and 256K (full) tokens of context

6 of its 52 layers use full attention and 46 are Mamba or feed-forward layers with no growing cache. Each extra token of context adds 6 KB of FP16 cache per request. The last column is the smallest setup here, one card or a group, that holds it at Q4_K_M. How the KV cache works.

How fast Nemotron 3 Nano 30B-A3B writes

Tokens per second for one request with 8,192 tokens of context, estimated from memory bandwidth and the parameters read per token. A dash means it does not fit on one card. Try other settings in the speed calculator.

Longest context on one GPU

How many tokens of context Nemotron 3 Nano 30B-A3B fits on a single card with one request, an FP16 KV cache and 0.5 GB left free. "Full" means the model's whole context window fits.

Model details

Parameters

31.6B (31,577,937,344)

Experts

128 routed experts, all loaded

Active per token

3.5B

Layers

6 of its 52 layers use full attention and 46 are Mamba or feed-forward layers with no growing cache

Attention cache

2 KV heads × 128

Context length

262,144 tokens

Published weights

58.8 GB (BF16)

On Hugging Face

nvidia/NVIDIA-Nemotron-3-Nano-30B-A3B-BF16

Reported benchmarks

GPQA (no tools) 73.0 · MMLU-Pro 78.3 · LiveCodeBench (v6 2025-08–2025-05) 68.3 (reasoning on). Vendor-reported: from the maker’s own model card, not rerun here; read 2026-09-30.

Nemotron 3 Nano 30B-A3B next to similar-sized models

The three models here closest to it in parameter count, at Q4_K_M with 8,192 tokens of context.