NVIDIA-Nemotron-3-Nano-Omni-30B-A3B-Reasoning GGUF size and VRAM requirements

License: other ⬇ 14,525 ❤ 134
Parameters31.58B
Context1,048,576

unsloth/NVIDIA-Nemotron-3-Nano-Omni-30B-A3B-Reasoning-GGUF is a very large reasoning-focused model with 31.58 billion parameters, built on the nemotron_h_moe architecture. It is released under the other license and has been downloaded 14,525 times.

To run unsloth/NVIDIA-Nemotron-3-Nano-Omni-30B-A3B-Reasoning-GGUF locally at a 4,096-token context, its quantized versions need between 2.9 GB (F16, lowest quality) and 34.39 GB (Q8_K_XL, highest quality) of memory, weights plus KV cache and a system margin included.

For most users the best balance is F32, needing about 4.38 GB. That means unsloth/NVIDIA-Nemotron-3-Nano-Omni-30B-A3B-Reasoning-GGUF fits entirely in the VRAM of a 6 GB GPU or larger, running fully on the GPU.

→ Guide: How much VRAM do you need?

GGUF file size and memory by quantization

Compare real GGUF weight sizes, estimated KV cache and total memory for Q4, Q5, Q8 and every quantization published in this repository.

Quant.Bits QualityWeights KVTotal Speed~Verdict
F16 0.4 Very low 1.48 GB 0.62 GB 2.9 GB 270.5 t/s Fits in VRAM
BF16 0.4 Very low 1.48 GB 0.62 GB 2.91 GB 270.2 t/s Fits in VRAM
F32 0.8 Very low 2.95 GB 0.62 GB 4.38 GB 135.6 t/s Fits in VRAM
IQ2_M 4.69 Good 17.23 GB 0.62 GB 18.65 GB 2.9 t/s Offload
IQ2_XXS 4.69 Good 17.23 GB 0.62 GB 18.65 GB 2.9 t/s Offload
Q2_K_XL 4.69 Good 17.23 GB 0.62 GB 18.66 GB 2.9 t/s Offload
IQ3_S 4.77 Good 17.53 GB 0.62 GB 18.95 GB 2.9 t/s Offload
IQ3_XXS 4.93 Good 18.12 GB 0.62 GB 19.55 GB 2.8 t/s Offload
IQ4_NL 4.95 Good 18.19 GB 0.62 GB 19.62 GB 2.7 t/s Offload
IQ4_NL_XL 4.95 Good 18.19 GB 0.62 GB 19.62 GB 2.7 t/s Offload
IQ4_XS 4.95 Good 18.19 GB 0.62 GB 19.62 GB 2.7 t/s Offload
Q3_K_M 4.95 Good 18.19 GB 0.62 GB 19.62 GB 2.7 t/s Offload
Q3_K_XL 4.95 Good 18.19 GB 0.62 GB 19.62 GB 2.7 t/s Offload
GGUF 5.51 Very good 20.24 GB 0.62 GB 21.66 GB 2.5 t/s Offload
Q4_K_S 5.84 Very good 21.47 GB 0.62 GB 22.89 GB 2.3 t/s Offload
Q4_K_M 6.05 Very good 22.25 GB 0.62 GB 23.67 GB 2.2 t/s Offload
Q4_K_XL 6.06 Very good 22.28 GB 0.62 GB 23.71 GB 2.2 t/s Offload
Q5_K_S 6.28 Very good 23.1 GB 0.62 GB 24.53 GB Insufficient
Q5_K_M 7.35 Excellent 27.0 GB 0.62 GB 28.43 GB Insufficient
Q5_K_XL 7.4 Excellent 27.19 GB 0.62 GB 28.62 GB Insufficient
Q8_0 8.51 Excellent 31.28 GB 0.62 GB 32.7 GB Insufficient
Q6_K 8.51 Excellent 31.28 GB 0.62 GB 32.7 GB Insufficient
Q6_K_XL 8.51 Excellent 31.28 GB 0.62 GB 32.7 GB Insufficient
Q8_K_XL 8.97 Excellent 32.97 GB 0.62 GB 34.39 GB Insufficient

KV cache computed from the model's exact architecture. Speed is a rough estimate bounded by memory bandwidth.

Frequently asked questions

How much VRAM do you need to run unsloth/NVIDIA-Nemotron-3-Nano-Omni-30B-A3B-Reasoning-GGUF?

You need about 4.38 GB of VRAM to run unsloth/NVIDIA-Nemotron-3-Nano-Omni-30B-A3B-Reasoning-GGUF entirely on the GPU using the F32 quantization (at a 4,096-token context). Smaller quantizations lower the requirement at the cost of quality.

Can I run unsloth/NVIDIA-Nemotron-3-Nano-Omni-30B-A3B-Reasoning-GGUF on an 8 GB GPU?

Yes. With 8 GB of VRAM you can run unsloth/NVIDIA-Nemotron-3-Nano-Omni-30B-A3B-Reasoning-GGUF fully on the GPU using F32 (about 4.38 GB).

Can I run unsloth/NVIDIA-Nemotron-3-Nano-Omni-30B-A3B-Reasoning-GGUF on a 16 GB GPU?

Yes. With 16 GB of VRAM you can run unsloth/NVIDIA-Nemotron-3-Nano-Omni-30B-A3B-Reasoning-GGUF fully on the GPU using F32 (about 4.38 GB).

Can I run unsloth/NVIDIA-Nemotron-3-Nano-Omni-30B-A3B-Reasoning-GGUF on a 24 GB GPU?

Yes. With 24 GB of VRAM you can run unsloth/NVIDIA-Nemotron-3-Nano-Omni-30B-A3B-Reasoning-GGUF fully on the GPU using Q4_K_XL (about 23.71 GB).

What is the best quantization for unsloth/NVIDIA-Nemotron-3-Nano-Omni-30B-A3B-Reasoning-GGUF?

If memory allows, higher bits-per-weight means better quality. A common sweet spot is a Q4_K_M or Q5_K_M quantization, which keeps most of the quality while roughly halving the memory versus 8-bit. Pick the highest quantization that still fits in your VRAM.