Nemotron Nano 9B v2 has 8.89B parameters. In BF16 its weights alone take 16.6 GiB; quantized to 4-bit, about 6.18 GiB. Each 32K-token request adds 0.50 GiB of KV cache. Cheapest way to run it today: 1 × RTX A4000 (16 GB) at $0.072 per hour. It calls tools and has a thinking mode you can switch off.
Read every day from Hugging Face. Split files are added together; vision projectors are left out.
How many GPUs to run Nemotron Nano 9B v2
Fewest GPUs that fit the weights plus one 8K-token request. Data-center GPUs use vLLM (90% of memory usable), the others llama.cpp. Several GPUs split the model with tensor or pipeline parallelism.
GPU
BF16 / FP16
8-bit
4-bit
GeForce RTX 4090 24 GB
1
1
1
GeForce RTX 5090 32 GB
1
1
1
RTX PRO 6000 Blackwell 96 GB
1
1
1
Mac M5 Max 128 GB 96 GB usable
1
1
1
Ryzen AI Max+ 395 128 GB 96 GB usable
1
1
1
DGX Spark (GB10) 128 GB
1
1
1
NVIDIA H100 80 GB
1
1
1
NVIDIA H200 141 GB
1
1
1
NVIDIA B200 (HGX) 180 GB
1
1
1
8-bit: FP8 with vLLM (INT8 on GPUs without FP8), Q8_0 with llama.cpp. 4-bit: AWQ / GPTQ with vLLM, Q4_K_M with llama.cpp, both counted at the Q4_K_M size.
Cheapest way to run Nemotron Nano 9B v2 today
For each GPU: the fewest cards that fit Nemotron Nano 9B v2 with one 8K-token request, the most faithful weight format at that count, and the cheapest on-demand price (2026-11-06).
On-demand prices from AWS, Google Cloud, Azure, RunPod, Vast.ai, Oracle Cloud, OVHcloud, Scaleway, Verda, Akash and UpCloud, checked every hour. Serving many users needs more KV cache, so more memory: size it in the calculator. Some links are affiliate links: StudioTV may earn a commission, at no extra cost to you.
Reasoning. Thinking mode you can switch on or off. With vLLM or SGLang, set it per request: "chat_template_kwargs": {"enable_thinking": false} turns it off, true on. Reasoning tokens take context and KV cache like any other token: leave room for them.
KV cache: how memory grows with context
Architecture: 56 layers · 4 full attention (8 KV × 128) · 27 Mamba-2. Its KV cache is 14.0× smaller than if every layer used full attention, because only 4 of its 56 layers keep a KV cache; the others hold a small fixed-size state.
Context per request
BF16 cache
FP8 cache
8K tokens
0.13 GiB
0.06 GiB
32K tokens
0.50 GiB
0.25 GiB
128K tokens
2.00 GiB
1.00 GiB
Plus 137 MiB of recurrent state per request, whatever the context.
Nemotron Nano 9B v2 VRAM FAQ
How much VRAM does Nemotron Nano 9B v2 need?
In BF16 the weights alone take 16.6 GiB (18 GB). In FP8 that is 9.37 GiB, and about 6.18 GiB with 4-bit quantization (Q4_K_M). Each request then adds KV cache: 0.13 GiB at 8K tokens and 0.50 GiB at 32K (BF16 cache).
Can Nemotron Nano 9B v2 run on a single RTX 4090 (24 GB)?
Yes. In BF16 it fits on one RTX 4090 with an 8K-token context (16.8 GiB including the cache, with llama.cpp).
Can Nemotron Nano 9B v2 run on an RTX 5090 (32 GB)?
Yes. In BF16 (16.8 GiB with an 8K-token context) it fits on one RTX 5090, with llama.cpp.
Can Nemotron Nano 9B v2 run on a Mac?
Yes. In 4-bit (Q4_K_M with llama.cpp) it needs a Mac with 16 GB of unified memory or more, of which macOS gives the GPU about 11 GB, for one request with an 8K-token context. The same Mac also holds it in 8-bit (Q8_0).
Can Nemotron Nano 9B v2 run on a DGX Spark or a Ryzen AI Max+ 395 (128 GB)?
Yes. In BF16 (16.8 GiB with an 8K-token context) it fits on a DGX Spark (128 GB) and on a Ryzen AI Max+ 395 with 128 GB, whose GPU gets up to 96 GB, with llama.cpp and one request.
What is the cheapest way to run Nemotron Nano 9B v2?
On 2026-11-06, the cheapest on-demand setup is 1 × RTX A4000 (16 GB) with Q8_0 weights on Vast.ai, at $0.072 per hour (about $52.56 per month). Next: 1 × NVIDIA L4 (24 GB) with FP8 weights on Vast.ai, at $0.104 per hour. Sized for one 8K-token request; prices are checked every hour.
How many H100 GPUs does Nemotron Nano 9B v2 need?
With one request and an 8K-token context: 1 in BF16, 1 in FP8 and 1 in 4-bit (vLLM, 90% of memory usable). Serving many users at once needs more memory for their KV caches: the calculator sizes that for you.
Does Nemotron Nano 9B v2 support tool calling?
Yes, Nemotron Nano 9B v2 supports tool (function) calling: it can ask your application to run a function and use the result.
Does Nemotron Nano 9B v2 have a thinking mode?
Thinking mode you can switch on or off. With vLLM or SGLang, set it per request with "chat_template_kwargs": {"enable_thinking": false} (or true). Reasoning tokens count toward the context and the KV cache.
Estimates, not guarantees: computed from the official config.json with the same engine as the LLM VRAM Calculator. Real usage depends on your engine version and settings.
NVIDIA, United States · Released August 12, 2025