Hugging FaceA 253B chat model from NVIDIA.
- Rating
- 3.5
- Downloads
- 2.2K
- Likes
- 16
- Parameters
- 253B
- Context
- 128K
- Licence
- other
Too big for this Mac
Its smallest size needs 74 GB of memory, more than M4 · 16 GB can give it.
Ratings
3.5 out of 5, across people's blind votes on its answers.
LM Arena ratings of October 28, 2026 (CC BY 4.0): people vote between two answers without knowing which model wrote which. The bar ends at 1488, the best any open model has. The Arena tests the full model; a small size on a Mac answers a little worse.
Llama 3 1 Nemotron Ultra 253B V1 runs well from a Mac with an M2 Max and 96 GB of memory. 15 of 59 Mac configurations run it well.
Compare with
The nearest models in size.
- Solar Open2 250BUpstage · 576 downloadsToo big for this Mac · 60 GB
- Qwen3 235B A22B 2507Alibaba Qwen · 24K downloadsToo big for this Mac · 53 GB
- Qwen3 235B A22BAlibaba Qwen · 18K downloads · Arena 1375Too big for this Mac · 86 GB
- MiniMax M2.7MiniMax · 15K downloads · Arena 1415Too big for this Mac · 47 GB
Sizes
From bartowski/nvidia_Llama-3_1-Nemotron-Ultra-253B-v1-GGUF on Hugging Face. A smaller size is a smaller download and a little less accurate.
| Size | Download | Memory at 8K | Get |
|---|---|---|---|
| IQ1_S | 54 GB | 55 GBestimated | Open |
| IQ1_M | 59 GB | 60 GBestimated | Open |
| IQ2_XXS | 67 GB | 69 GBestimated | Open |
| IQ2_XS | 75 GB | 76 GBestimated | Open |
| IQ2_S | 79 GB | 80 GBestimated | Open |
| IQ2_M | 85 GB | 87 GBestimated | Open |
| Q2_K | 93 GB | 95 GBestimated | Open |
| IQ3_XXS | 98 GB | 99 GBestimated | Open |
| IQ3_XS | 103 GB | 105 GBestimated | Open |
| Q3_K_S | 110 GB | 111 GBestimated | Open |
| IQ3_M | 114 GB | 115 GBestimated | Open |
| Q3_K_M | 122 GB | 123 GBestimated | Open |
| Q3_K_L | 133 GB | 134 GBestimated | Open |
| Q3_K_XL | 135 GB | 136 GBestimated | Open |
| IQ4_XS | 135 GB | 137 GBestimated | Open |
| IQ4_NL | 143 GB | 145 GBestimated | Open |
| Q4_0 | 144 GB | 145 GBestimated | Open |
| Q4_K_S | 144 GB | 146 GBestimated | Open |
| Q4_K_M | 151 GB | 152 GBestimated | Open |
| Q4_K_L | 152 GB | 154 GBestimated | Open |
| Q4_1 | 159 GB | 160 GBestimated | Open |
| Q5_K_S | 175 GB | 176 GBestimated | Open |
| Q5_K_M | 179 GB | 180 GBestimated | Open |
| Q6_K | 208 GB | 209 GBestimated | Open |
| Q8_0 | 269 GB | 271 GBestimated | Open |
Run It Anywhere
Llama 3 1 Nemotron Ultra 253B V1 is an open model, so any app that runs GGUF files can run it. These take the IQ1_S size from bartowski straight from Hugging Face.
- LM StudioMac, Windows, Linux
Open - JanMac, Windows, Linux
Open - MstyMac, Windows, Linux
Open - OllamaMac, Windows, Linux
ollama run hf.co/bartowski/nvidia_Llama-3_1-Nemotron-Ultra-253B-v1-GGUF:IQ1_S - llama.cppMac, Windows, Linux
llama-server -hf bartowski/nvidia_Llama-3_1-Nemotron-Ultra-253B-v1-GGUF:IQ1_S - Docker Model RunnerMac, Windows, Linux
docker model run hf.co/bartowski/nvidia_Llama-3_1-Nemotron-Ultra-253B-v1-GGUF:IQ1_S
On a PC Graphics Card
| Video memory | Largest size that fits | Download | Memory at 8K |
|---|---|---|---|
| 6 GB | None worth running | needs 99 GB | |
| 8 GB | None worth running | needs 99 GB | |
| 10 GB | None worth running | needs 99 GB | |
| 11 GB | None worth running | needs 99 GB | |
| 12 GB | None worth running | needs 99 GB | |
| 16 GB | None worth running | needs 99 GB | |
| 20 GB | None worth running | needs 99 GB | |
| 24 GB | None worth running | needs 99 GB | |
| 32 GB | None worth running | needs 99 GB | |
| 48 GB | None worth running | needs 99 GB |
The largest size, of those that keep three bits a weight or more, whose memory at an 8K chat is under 92% of the card's; the rest is the desktop's. Estimated from the files, the same way as for a Mac; speed on a PC is not estimated yet. A size that doesn't fit the card still runs, slower, with part of it in ordinary memory.
On every Mac
| Mac | Size | Fit | Memory | Writes |
|---|---|---|---|---|
| M6 · 16 GB | IQ1_S | Too big | 74 GB | |
| M6 · 24 GB | IQ1_S | Too big | 74 GB | |
| M6 · 32 GB | IQ1_S | Too big | 74 GB | |
| M5 · 16 GB | IQ1_S | Too big | 74 GB | |
| M5 · 24 GB | IQ1_S | Too big | 74 GB | |
| M5 · 32 GB | IQ1_S | Too big | 74 GB | |
| M5 Pro · 24 GB | IQ1_S | Too big | 74 GB | |
| M5 Pro · 48 GB | IQ1_S | Too big | 74 GB | |
| M5 Pro · 64 GB | IQ1_S | Too big | 74 GB | |
| M5 Max · 36 GB | IQ1_S | Too big | 74 GB | |
| M5 Max · 48 GB | IQ1_S | Too big | 74 GB | |
| M5 Max · 64 GB | IQ1_S | Too big | 74 GB | |
| M5 Max · 128 GB | IQ3_XXS | Runs well | 100 GB | 3.2 |
| M5 Ultra · 96 GB | UD-IQ1_M | Runs well | 65 GB | 9.3 |
| M5 Ultra · 256 GB | Q4_1 | Runs great | 169 GB | 4.6 |
| M5 Ultra · 512 GB | Q8_0 | Runs great | 289 GB | 2.9 |
| M4 · 16 GB | IQ1_S | Too big | 74 GB | |
| M4 · 24 GB | IQ1_S | Too big | 74 GB | |
| M4 · 32 GB | IQ1_S | Too big | 74 GB | |
| M4 Pro · 24 GB | IQ1_S | Too big | 74 GB | |
| M4 Pro · 48 GB | IQ1_S | Too big | 74 GB | |
| M4 Pro · 64 GB | IQ1_S | Too big | 74 GB | |
| M4 Max · 36 GB | IQ1_S | Too big | 74 GB | |
| M4 Max · 48 GB | IQ1_S | Too big | 74 GB | |
| M4 Max · 64 GB | IQ1_S | Too big | 74 GB | |
| M4 Max · 96 GB | UD-IQ1_M | Runs well | 65 GB | 3.1 |
| M4 Max · 128 GB | IQ3_XXS | Runs well | 100 GB | 2.7 |
| M3 · 8 GB | IQ1_S | Too big | 74 GB | |
| M3 · 16 GB | IQ1_S | Too big | 74 GB | |
| M3 · 24 GB | IQ1_S | Too big | 74 GB | |
| M3 Pro · 18 GB | IQ1_S | Too big | 74 GB | |
| M3 Pro · 36 GB | IQ1_S | Too big | 74 GB | |
| M3 Max · 36 GB | IQ1_S | Too big | 74 GB | |
| M3 Max · 48 GB | IQ1_S | Too big | 74 GB | |
| M3 Max · 64 GB | IQ1_S | Too big | 74 GB | |
| M3 Max · 96 GB | UD-IQ1_M | Runs well | 65 GB | 2.4 |
| M3 Max · 128 GB | IQ3_XXS | Runs well | 100 GB | 2.1 |
| M3 Ultra · 96 GB | UD-IQ1_M | Runs well | 65 GB | 5.7 |
| M3 Ultra · 256 GB | Q4_1 | Runs great | 169 GB | 3 |
| M3 Ultra · 512 GB | Q8_0 | Runs great | 289 GB | 1.8 |
| M2 · 8 GB | IQ1_S | Too big | 74 GB | |
| M2 · 16 GB | IQ1_S | Too big | 74 GB | |
| M2 · 24 GB | IQ1_S | Too big | 74 GB | |
| M2 Pro · 16 GB | IQ1_S | Too big | 74 GB | |
| M2 Pro · 32 GB | IQ1_S | Too big | 74 GB | |
| M2 Max · 32 GB | IQ1_S | Too big | 74 GB | |
| M2 Max · 64 GB | IQ1_S | Too big | 74 GB | |
| M2 Max · 96 GB | UD-IQ1_M | Runs well | 65 GB | 3.1 |
| M2 Ultra · 64 GB | IQ1_S | Too big | 74 GB | |
| M2 Ultra · 128 GB | IQ3_XXS | Runs well | 100 GB | 3.9 |
| M2 Ultra · 192 GB | Q4_K_M | Runs well | 152 GB | 2.7 |
| M1 · 8 GB | IQ1_S | Too big | 74 GB | |
| M1 · 16 GB | IQ1_S | Too big | 74 GB | |
| M1 Pro · 16 GB | IQ1_S | Too big | 74 GB | |
| M1 Pro · 32 GB | IQ1_S | Too big | 74 GB | |
| M1 Max · 32 GB | IQ1_S | Too big | 74 GB | |
| M1 Max · 64 GB | IQ1_S | Too big | 74 GB | |
| M1 Ultra · 64 GB | IQ1_S | Too big | 74 GB | |
| M1 Ultra · 128 GB | IQ3_XXS | Runs well | 100 GB | 3.6 |
Writes is words a second, estimated from the Mac's memory speed and the model's size; how, and how wrong.
Packaged by
3 repositories on Hugging Face carry this model for llama.cpp. Outmake reads them all as one model.
Information
- Maker
- NVIDIA
- Released
- April 2025
- Parameters
- 253B
- Longest chat
- 128K tokens
- Reads pictures
- No
- Licence
- other
- Downloads this month
- 2,237
- Likes
- 16
- Sizes
- 25, 54 to 269 GB
- Kind
- Maker's release
Badge for the model card
Shows on Llama 3 1 Nemotron Ultra 253B V1's Hugging Face card which Macs run it well, linked to this page. It updates every morning.
Files, downloads and likes read from Hugging Face on November 5, 2026 and refreshed every morning. Memory and speed are estimates until someone running Outmake measures this model on that Mac.