NVIDIA Nemotron 3 Nano 30B A3B FP8 — Hardware Requirements & GPU Compatibility
Chat
NVIDIA Nemotron 3 Nano 30B A3B FP8 is the FP8-quantized version of NVIDIA's 31.6 billion parameter mixture-of-experts model. The 8-bit floating point format reduces memory requirements compared to BF16 while retaining strong output quality, making it a practical choice for GPUs with tighter VRAM budgets. With only about 3 billion parameters active per token, this model already runs efficiently. The FP8 quantization pushes the memory savings further without meaningful degradation, making it one of the best options for users who want MoE-class performance on mainstream hardware.
1.5M downloads
305 likesSpecifications
- Publisher
- NVIDIA
- Parameters
- 31.6B
- Architecture
- NemotronHForCausalLM
- Context Length
- 262,144 tokens
- Vocabulary Size
- 131,072
- Release Date
- 2026-03-15
- License
- Other
Get Started
HuggingFace
How Much VRAM Does NVIDIA Nemotron 3 Nano 30B A3B FP8 Need?
Select a quantization to see compatible GPUs below.
| Quantization | Bits | VRAM | + Context | File Size | Quality |
|---|---|---|---|---|---|
| BF16 | 16.00 | 63.5 GB | 72.6 GB | 63.16 GB | Brain floating point 16 — preferred for training |
Which GPUs Can Run NVIDIA Nemotron 3 Nano 30B A3B FP8?
BF16 · 63.5 GBNVIDIA Nemotron 3 Nano 30B A3B FP8 (BF16) requires 63.5 GB of VRAM to load the model weights. For comfortable inference with headroom for KV cache and system overhead, 83+ GB is recommended. Using the full 262K context window can add up to 9.1 GB, bringing total usage to 72.6 GB. No single GPU has enough memory — multi-GPU or cluster setups are needed.
Which Devices Can Run NVIDIA Nemotron 3 Nano 30B A3B FP8?
BF16 · 63.5 GB8 devices with unified memory can run NVIDIA Nemotron 3 Nano 30B A3B FP8, including NVIDIA DGX H100, NVIDIA DGX A100 640GB, Mac Studio M4 Max (64 GB).
Runs great
— Plenty of headroom
Decent
— Enough memory, may be tightRelated Models
Frequently Asked Questions
- How much VRAM does NVIDIA Nemotron 3 Nano 30B A3B FP8 need?
NVIDIA Nemotron 3 Nano 30B A3B FP8 requires 63.5 GB of VRAM at BF16. Full 262K context adds up to 9.1 GB (72.6 GB total).
VRAM = Weights + KV Cache + Overhead
Weights = 31.6B × 16 bits ÷ 8 = 63.2 GB
KV Cache + Overhead ≈ 0.3 GB (at 2K context + ~0.3 GB framework)
KV Cache + Overhead ≈ 9.4 GB (at full 262K context)
VRAM usage by quantization
BF1663.5 GBBF16 + full context72.6 GB- Can NVIDIA GeForce RTX 5090 run NVIDIA Nemotron 3 Nano 30B A3B FP8?
No — NVIDIA Nemotron 3 Nano 30B A3B FP8 requires at least 63.5 GB at BF16, which exceeds the NVIDIA GeForce RTX 5090's 32 GB of VRAM.
- Can I run NVIDIA Nemotron 3 Nano 30B A3B FP8 on a Mac?
NVIDIA Nemotron 3 Nano 30B A3B FP8 requires at least 63.5 GB at BF16, which exceeds the unified memory of most consumer Macs. You would need a Mac Studio or Mac Pro with a high-memory configuration.
- Can I run NVIDIA Nemotron 3 Nano 30B A3B FP8 locally?
Yes — NVIDIA Nemotron 3 Nano 30B A3B FP8 can run locally on consumer hardware. At BF16 quantization it needs 63.5 GB of VRAM. Popular tools include Ollama, LM Studio, and llama.cpp.
- How fast is NVIDIA Nemotron 3 Nano 30B A3B FP8?
At BF16, NVIDIA Nemotron 3 Nano 30B A3B FP8 can reach ~46 tok/s on AMD Instinct MI300X. Speed depends mainly on GPU memory bandwidth. Real-world results typically within ±20%.
tok/s = (bandwidth GB/s ÷ model GB) × efficiency
Example: AMD Instinct MI300X → 5300 ÷ 63.5 × 0.55 = ~46 tok/s
Estimated speed at BF16 (63.5 GB)
AMD Instinct MI300X~46 tok/sNVIDIA H100 SXM~34 tok/sAMD Instinct MI250X~28 tok/sReal-world results typically within ±20%. Speed depends on batch size, quantization kernel, and software stack.
- What's the download size of NVIDIA Nemotron 3 Nano 30B A3B FP8?
At BF16, the download is about 63.16 GB.