NVIDIA Nemotron 3 Super 120B A12B BF16 Heretic — Hardware Requirements & GPU Compatibility
Chat
NVIDIA Nemotron 3 Super 120B A12B BF16 Heretic is a 120.7B-parameter open language model from trohrbaugh in the Nemotron family. It supports a context window of up to 262,144 tokens. At Q4_K_M it needs about 72.89 GB of VRAM — see which GPUs and Macs can run it below.
NVIDIA Nemotron 3 Super 120B A12B BF16 Heretic runs on a 96 GB+ Mac or a workstation GPU
Best local pick → Mac Studio (M3 Ultra, 96GB) · Q4_K_M · 72.9 GB · ~35 tok/s
Macs from 64 GB (Q3_K_M · ~38 tok/s)
Other
Not set up for it?
Reference GPU · RTX 4090 24 GB
Estimates, ±20% · how we estimate
Run NVIDIA Nemotron 3 Super 120B A12B BF16 Heretic
One recommended way. Every other method, OS and quant is in the step-by-step guide.Each way opens a step-by-step guide for your OS and quant.
Recommended
ollama run hf.co/mradermacher/NVIDIA-Nemotron-3-Super-120B-A12B-BF16-heretic-i1-GGUFQ4_K_M · needs ~72.9 GB · 16K context
Download: NVIDIA-Nemotron-3-Super-120B-A12B-BF16-heretic.i1-Q4_K_M.gguf86.05 GBmradermacher
Ways to runOther ways
each opens step-by-stepRun in cloud
Runs on RTX PRO 6000 (96 GB) (23 GB headroom) · Q4_K_M
- Generation speed
- ~99 tok/s
- generation speed
- Cost per 1M output tokens
- $3.72
- per 1M output tokens
How Much VRAM Does NVIDIA Nemotron 3 Super 120B A12B BF16 Heretic Need?
Select a quantization level; the listed hardware adheres to it.
| Quantization | VRAM | +32K | On RTX 4090 (reference) | |
|---|---|---|---|---|
| 51.8 GB | 52 GB | 2-bit · K-quant — 2-bit quantization with K-quant improvements | Does not fit: needs 51.8 GB | |
| 59.3 GB | 59.5 GB | 3-bit medium — 3-bit medium quantization | Does not fit: needs 59.3 GB | |
| 72.9 GB | 73.1 GB | 4-bit medium — 4-bit medium quantization — most popular sweet spot | Does not fit: needs 72.9 GB | |
| 86.5 GB | 86.6 GB | 5-bit medium — 5-bit medium quantization — good quality/size tradeoff | Does not fit: needs 86.5 GB | |
| 100.0 GB | 100.2 GB | 6-bit, very good — 6-bit quantization, very good quality | Does not fit: needs 100 GB | |
| 121.2 GB | 121.3 GB | 8-bit, near-lossless — 8-bit quantization, near-lossless | Does not fit: needs 121.2 GB | |
| 241.8 GB | 242 GB | 16-bit · bfloat — Brain floating point 16 — preferred for training | Does not fit: needs 241.8 GB |
Selected Q4_K_M
4-bit medium quantization72.9GBat 16K context
- Weights
- 72.4 GB
- KV cache
- 0.2 GB
- Overhead
- 0.3 GB
Doesn't fit RTX 4090 (reference) · needs 48.9 GB more
Or .
Which Hardware Can Run NVIDIA Nemotron 3 Super 120B A12B BF16 Heretic?
Q4_K_M · 72.9 GBNVIDIA Nemotron 3 Super 120B A12B BF16 Heretic (Q4_K_M) requires 72.9 GB of VRAM to load the model weights. For comfortable inference with headroom for KV cache and system overhead, 95+ GB is recommended. Using the full 262K context window can add up to 2.8 GB, bringing total usage to 75.7 GB. No consumer GPU has enough memory. 30 devices with unified memory can run NVIDIA Nemotron 3 Super 120B A12B BF16 Heretic, including Mac Studio (M3 Ultra, 256GB), Mac Studio (M3 Ultra, 512GB), Mac Studio (M3 Ultra, 96GB).
64–80 GB
Q4_K_M · best fit Q3_K_M (59.3 GB)
3 Mac classes · 1 device family
M2 Ultra*64 GB~38Simulate output on Mac Pro M2 Ultra (64 GB)M5 Max64 GB~34Simulate output on Mac Studio M5 Max (64 GB)Orin~24Simulate output on NVIDIA Jetson AGX Orin 64GB96 GB
fits · 23.1 GB free · up to 256K ctx
2 Mac classes
M3 Ultra*96 GB~35Simulate output on Mac Studio (M3 Ultra, 96GB)M3 Max*96 GB~22Simulate output on MacBook Pro 14" M3 Max (96 GB)
By VRAM tier · typical GPUs and Macs at Q4_K_M
Alternatives to NVIDIA Nemotron 3 Super 120B A12B BF16 Heretic
Nemotron family · 45 models →Similar models
6| Model | Score | VRAM | tok/s | Why it is similar |
|---|---|---|---|---|
| Qwen3.5 122B A10BAlibaba | 109.4 | 75.5 GB | —needs > 24 GB | same size |
| GPT OSS 120BOpenAI | 97.5 | 70.8 GB | —needs > 24 GB | same size |
| Ornith 1.5 122Bornith-ai | — | 75.5 GB | —needs > 24 GB | same size |
| Ling 3.0 Flash VLInclusion AI | — | 76 GB | —needs > 24 GB | same size |
| Leanstral 1.5 119B A6BMistral AI | — | 84 GB | —needs > 24 GB | same size |
| Laguna S 2.1poolside | — | 72.5 GB | —needs > 24 GB | same size |
Specifications
- Publisher
- trohrbaugh
- Family
- Nemotron
- Parameters
- 120.7B
- Active parameters
- 10.1B per token
- Architecture
- NemotronHForCausalLM
- Context Length
- 262,144 tokens
- Vocabulary Size
- 131,072
- Release Date
- 2026-03-15
- License
- Other
Frequently Asked Questions
How much VRAM does NVIDIA Nemotron 3 Super 120B A12B BF16 Heretic need?
NVIDIA Nemotron 3 Super 120B A12B BF16 Heretic requires 72.9 GB of VRAM at Q4_K_M, or 241.8 GB at BF16. Full 262K context adds up to 2.8 GB (75.7 GB total).
VRAM = Weights + KV Cache + Overhead
Weights = 120.7B × 4.8 bits ÷ 8 = 72.4 GB
KV Cache + Overhead ≈ 0.5 GB (at 2K context + ~0.3 GB framework)
Fit ratings and hardware model lists check this model with room for a 16K-token context, which needs a little more memory.
KV Cache + Overhead ≈ 3.3 GB (at full 262K context)
VRAM usage by quantization
Can NVIDIA GeForce RTX 5090 run NVIDIA Nemotron 3 Super 120B A12B BF16 Heretic?
No — NVIDIA Nemotron 3 Super 120B A12B BF16 Heretic requires at least 51.8 GB at Q2_K, which exceeds the NVIDIA GeForce RTX 5090's 32 GB of VRAM.
What's the best quantization for NVIDIA Nemotron 3 Super 120B A12B BF16 Heretic?
For NVIDIA Nemotron 3 Super 120B A12B BF16 Heretic, Q4_K_M (72.9 GB) offers the best balance of quality and VRAM usage. Q5_K_M (86.5 GB) provides better quality if you have the VRAM. The smallest option is Q2_K at 51.8 GB.
VRAM requirement by quantization
★ Recommended — best balance of quality and VRAM usage.
Can I run NVIDIA Nemotron 3 Super 120B A12B BF16 Heretic on a Mac?
Yes — Mac Pro M2 Ultra (64 GB) and 41 other Macs can run NVIDIA Nemotron 3 Super 120B A12B BF16 Heretic, the smallest at Q3_K_M. Apple Silicon uses unified memory, so the model shares RAM with the system. At Q4_K_M you need at least 72.9 GB of usable unified memory (RAM minus macOS overhead), from Mac Studio (M3 Ultra, 96GB).
Can I run NVIDIA Nemotron 3 Super 120B A12B BF16 Heretic locally?
Yes — NVIDIA Nemotron 3 Super 120B A12B BF16 Heretic can run locally on consumer hardware. Popular tools include Ollama, LM Studio, and llama.cpp.
How fast is NVIDIA Nemotron 3 Super 120B A12B BF16 Heretic?
At Q4_K_M, NVIDIA Nemotron 3 Super 120B A12B BF16 Heretic can reach ~53 tok/s on AMD Instinct MI350X. Speed depends mainly on GPU memory bandwidth. Real-world results typically within ±20%.
tok/s = 1000 ÷ (active GB ÷ (bandwidth GB/s × efficiency) × 1000 + layers × routing ms)
Mixture-of-Experts: just the active experts get read for each token, along with a fixed routing cost per layer.
Example: NVIDIA B200 → 6.1 GB active ÷ (8000 × 0.65) = 1.17 ms, plus 88 layers × 0.055 ms = 4.84 ms, so 1000 ÷ 6.01 ms = ~166 tok/s
Estimated speed at Q4_K_M (72.9 GB)
Real-world results typically within ±20%. Speed depends on batch size, quantization kernel, and software stack.
What's the download size of NVIDIA Nemotron 3 Super 120B A12B BF16 Heretic?
At Q4_K_M, the download is about 72.40 GB. The full-precision BF16 version is 241.34 GB. The smallest option (Q2_K) is 51.28 GB.
Which GPUs can run NVIDIA Nemotron 3 Super 120B A12B BF16 Heretic?
No single consumer GPU has enough VRAM to run NVIDIA Nemotron 3 Super 120B A12B BF16 Heretic at Q4_K_M (72.9 GB). Multi-GPU or professional hardware is required.
Which devices can run NVIDIA Nemotron 3 Super 120B A12B BF16 Heretic?
30 devices with unified memory can run NVIDIA Nemotron 3 Super 120B A12B BF16 Heretic at Q4_K_M (72.9 GB), including ASUS Ascent GX10, Beelink GTR9 Pro (Ryzen AI Max+ 395, 128 GB), Framework Desktop (Ryzen AI Max+ 395, 128 GB), GMKtec EVO-X2 (Ryzen AI Max+ 395, 128 GB). Apple Silicon Macs use unified memory shared between CPU and GPU, making them well-suited for local LLM inference.