Llama 3.1 Nemotron Ultra 253B v1 on NVIDIA H100 NVL

Can the NVIDIA H100 NVL run Llama 3.1 Nemotron Ultra 253B v1?

Barely: it runs very slowly (offloads to RAM)

Very slow (offloads to RAM). At Q2_K it needs about 111 GB of 94.0 GB usable, with an estimated speed of 2.9 tok/s.

Best quant: Q2_K
Generation speed: 2.9 tok/s
Prompt processing: 301 tok/s
First token (1K prompt): 3.4 s
Largest safe context: —

Best app to run it on the NVIDIA H100 NVL

On Linux

LM Studio

Desktop app · runs on llama.cpp, MLX

Splits the model between GPU memory and system RAM, which this setup needs.
Point-and-click: search, download and chat in one window.

Open Llama 3.1 Nemotron Ultra 253B v1 in LM Studio

llama.cpp Splits the model between GPU memory and system RAM, which this setup needs.
Ollama Splits the model between GPU memory and system RAM, which this setup needs.

On Windows

LM Studio

Desktop app · runs on llama.cpp, MLX

Splits the model between GPU memory and system RAM, which this setup needs.
Point-and-click: search, download and chat in one window.

Open Llama 3.1 Nemotron Ultra 253B v1 in LM Studio

llama.cpp Splits the model between GPU memory and system RAM, which this setup needs.
Ollama Splits the model between GPU memory and system RAM, which this setup needs.

Want an agent? Llama 3.1 Nemotron Ultra 253B v1 supports tool use, so you can pair a llama.cpp or MLX LM server with one of these. Apps that run on this computer come first.

Coding agents

Personal assistants

How we choose: the files published for this model, your operating system and GPU, whether it fits fully in memory, and published speed tests (MLX vs llama.cpp on Apple silicon, vLLM vs llama.cpp vs Ollama on NVIDIA). Details.

Download Llama 3.1 Nemotron Ultra 253B v1

Recommended for the NVIDIA H100 NVL: Q2_K. Every other quantization has its own download link in the table below.

Quantized files · GGUF

Works in LM Studio, llama.cpp, Jan and Ollama. No sign-in needed.

Download the Q2_K files (opens in a new tab)

huggingface.co/unsloth/Llama-3_1-Nemotron-Ultra-253B-v1-GGUF (opens in a new tab)

ollama run hf.co/unsloth/Llama-3_1-Nemotron-Ultra-253B-v1-GGUF

Original weights · Hugging Face

Full-precision files from NVIDIA, for fine-tuning, vLLM, Transformers.

huggingface.co/nvidia/Llama-3_1-Nemotron-Ultra-253B-v1 (opens in a new tab)

No sign-in needed.

Links checked 3 days ago against Hugging Face, Ollama and GitHub.

How it’s rated

Quality 60Good quality (B) · 60 of 100: how capable the model is at Q2_K, whatever it runs on. Match score 56/100 (quality, fit, speed and memory use) sets the ranking.: How capable the model itself is at Q2_K, from its editorial quality tier scaled by the quantization or precision. It’s the same on any hardware that runs this variant. Scale: A+ 85+ · A 70+ · B 55+ · C 40+ · D under 40.
Runs Very slowoffloads to RAM: How well it runs on the NVIDIA H100 NVL: speed, capped by how well it fits in memory. Runs great: 30+ tokens/s (or near the image, video or decision speed target) · Runs well: 10+ tokens/s · Slow: 3+ tokens/s · Very slow: under 3 · Won’t run: doesn’t fit even with offload. A tight fit or RAM offload caps it at Runs well; heavy offload caps it at Slow.

Match score

To order lists, each result also gets a match score out of 100 from four parts. It favors the most capable model that still runs well.

Model qualityEditorial quality tier of the model, scaled by the quantization or precision used.	27.1 / 45
Memory fitFull marks when everything sits in GPU memory; fewer when part spills to system RAM.	10.0 / 25
SpeedGeneration speed against a 40 tokens/s target for LLMs, or render time against a target for image and video.	4.1 / 15
Memory useRewards using the memory you have, so a bigger, better variant beats a tiny one that leaves most memory idle.	15.0 / 15
Match score	56.2 / 100

The match score orders results; it is not shown as a grade. Quality and Runs are the two ratings in the tables.

Every quantization on the NVIDIA H100 NVL

Quant	Memory needed	Available	Speed	Runs	Quality	Download
`Q2_K`	111 GB	94.0 GB	2.9 tok/s	Very slowoffloads to RAM	60Good quality (B) · 60 of 100: how capable the model is at Q2_K, whatever it runs on. Match score 56/100 (quality, fit, speed and memory use) sets the ranking.	GGUF · folder (opens in a new tab)
`Q3_K_M`	134 GB	94.0 GB	—	Won't run	72Great quality (A) · 72 of 100: how capable the model is at Q3_K_M, whatever it runs on. Match score 9/100 (quality, fit, speed and memory use) sets the ranking.	GGUF · folder (opens in a new tab)
`Q4_K_M`	164 GB	94.0 GB	—	Won't run	81Great quality (A) · 81 of 100: how capable the model is at Q4_K_M, whatever it runs on. Match score 11/100 (quality, fit, speed and memory use) sets the ranking.	GGUF · folder (opens in a new tab)
`Q5_K_M`	189 GB	94.0 GB	—	Won't run	83Great quality (A) · 83 of 100: how capable the model is at Q5_K_M, whatever it runs on. Match score 11/100 (quality, fit, speed and memory use) sets the ranking.	GGUF · folder (opens in a new tab)
`Q6_K`	217 GB	94.0 GB	—	Won't run	85Excellent quality (A+) · 85 of 100: how capable the model is at Q6_K, whatever it runs on. Match score 11/100 (quality, fit, speed and memory use) sets the ranking.	GGUF · folder (opens in a new tab)
`Q8_0`	278 GB	94.0 GB	—	Won't run	86Excellent quality (A+) · 86 of 100: how capable the model is at Q8_0, whatever it runs on. Match score 11/100 (quality, fit, speed and memory use) sets the ranking.	GGUF · folder (opens in a new tab)
`F16`	516 GB	94.0 GB	—	Won't run	86Excellent quality (A+) · 86 of 100: how capable the model is at F16, whatever it runs on. Match score 11/100 (quality, fit, speed and memory use) sets the ranking.	GGUF · folder (opens in a new tab) Original weights (opens in a new tab)

Memory figures assume 8K of context. Longer conversations or documents grow the KV cache.

Other hardware that runs it

81Great quality (A) · 81 of 100: how capable the model is at Q4_K_M, whatever it runs on. Match score 90/100 (quality, fit, speed and memory use) sets the ranking. NVIDIA GB300 Grace Blackwell Ultra Desktop Superchip Q4_K_M · 34.4 tok/s · Runs great
81Great quality (A) · 81 of 100: how capable the model is at Q4_K_M, whatever it runs on. Match score 90/100 (quality, fit, speed and memory use) sets the ranking. NVIDIA B200 Q4_K_M · 38.8 tok/s · Runs well (tight fit)
81Great quality (A) · 81 of 100: how capable the model is at Q4_K_M, whatever it runs on. Match score 89/100 (quality, fit, speed and memory use) sets the ranking. AMD Instinct MI350X Q4_K_M · 33.4 tok/s · Runs great
81Great quality (A) · 81 of 100: how capable the model is at Q4_K_M, whatever it runs on. Match score 88/100 (quality, fit, speed and memory use) sets the ranking. AMD Instinct MI325X Q4_K_M · 25.0 tok/s · Runs well
81Great quality (A) · 81 of 100: how capable the model is at Q4_K_M, whatever it runs on. Match score 86/100 (quality, fit, speed and memory use) sets the ranking. AMD Instinct MI300X Q4_K_M · 22.1 tok/s · Runs well (tight fit)
72Great quality (A) · 72 of 100: how capable the model is at Q3_K_M, whatever it runs on. Match score 84/100 (quality, fit, speed and memory use) sets the ranking. NVIDIA H200 141GB SXM Q3_K_M · 30.1 tok/s · Runs well (tight fit)

Quick answers

How much memory does Llama 3.1 Nemotron Ultra 253B v1 need on the NVIDIA H100 NVL 94GB?

At Q2_K it needs about 111 GB. The NVIDIA H100 NVL 94GB has 94.0 GB usable.

How fast is Llama 3.1 Nemotron Ultra 253B v1 on the NVIDIA H100 NVL 94GB?

About 2.9 tok/s at Q2_K, an estimate from the hardware’s memory bandwidth and the size of the model.

Which quantization of Llama 3.1 Nemotron Ultra 253B v1 should I use on the NVIDIA H100 NVL 94GB?

Q2_K. Of the versions that fit, it is the one that runs best; when two run equally well, the check prefers the higher-rated one.

Estimates come from published hardware specs and our open formulas, not from running the model on your machine. Real speed varies with runtime, drivers and settings. Numbers marked “Measured” come from a linked benchmark.