Hugging FaceA 50B chat model from NVIDIA.
- Rating
- 3.4
- Downloads
- 3.4K
- Likes
- 11
- Parameters
- 50B
- Context
- 128K
- Licence
- other
Too big for this Mac
Its smallest size needs 27 GB of memory, more than M4 · 16 GB can give it.
Ratings
3.4 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 3 Nemotron Super 49B V1 5 runs well from a Mac with an M2 and 24 GB of memory. 47 of 59 Mac configurations run it well.
Compare with
The nearest models in size.
- Llama 3 3 Nemotron Super 49B V1NVIDIA · 2.7K downloads · Arena 1328Too big for this Mac · 14 GB
- Llama 3 1 Nemotron 51BNVIDIA · 3.1K downloads · Arena 1287Too big for this Mac · 14 GB
- Mixtral 8x7B V0.1Mistral AI · 41K downloads · Arena 1197Too big for this Mac · 16 GB
- FlintRecursiveSI · 469K downloadsTight fit · 12 GB · 19 words/s
Sizes
From unsloth/Llama-3_3-Nemotron-Super-49B-v1_5-GGUF on Hugging Face. A smaller size is a smaller download and a little less accurate.
| Size | Download | Memory at 8K | Get |
|---|---|---|---|
| UD-IQ1_S | 11 GB | 13 GBestimated | Open |
| UD-IQ1_M | 12 GB | 13 GBestimated | Open |
| UD-IQ2_XXS | 14 GB | 15 GBestimated | Open |
| UD-IQ2_M | 17 GB | 18 GBestimated | Open |
| Q2_K | 19 GB | 20 GBestimated | Open |
| Q2_K_L | 19 GB | 20 GBestimated | Open |
| UD-Q2_K_XL | 19 GB | 20 GBestimated | Open |
| UD-IQ3_XXS | 20 GB | 21 GBestimated | Open |
| Q3_K_S | 22 GB | 23 GBestimated | Open |
| Q3_K_M | 24 GB | 25 GBestimated | Open |
| UD-Q3_K_XL | 25 GB | 26 GBestimated | Open |
| IQ4_XS | 27 GB | 28 GBestimated | Open |
| IQ4_NL | 28 GB | 29 GBestimated | Open |
| Q4_0 | 28 GB | 30 GBestimated | Open |
| Q4_K_S | 29 GB | 30 GBestimated | Open |
| Q4_K_M | 30 GB | 31 GBestimated | Open |
| UD-Q4_K_XL | 30 GB | 31 GBestimated | Open |
| Q4_1 | 31 GB | 32 GBestimated | Open |
| Q5_K_S | 34 GB | 35 GBestimated | Open |
| Q5_K_M | 35 GB | 36 GBestimated | Open |
| UD-Q5_K_XL | 35 GB | 36 GBestimated | Open |
| Q6_K | 41 GB | 42 GBestimated | Open |
| UD-Q6_K_XL | 43 GB | 44 GBestimated | Open |
| Q8_0 | 53 GB | 54 GBestimated | Open |
| UD-Q8_K_XL | 59 GB | 60 GBestimated | Open |
| BF16 | 100 GB | 101 GBestimated | Open |
Run It Anywhere
Llama 3 3 Nemotron Super 49B V1 5 is an open model, so any app that runs GGUF files can run it. These take the UD-IQ1_S size from unsloth straight from Hugging Face.
- LM StudioMac, Windows, Linux
Open - JanMac, Windows, Linux
Open - MstyMac, Windows, Linux
Open - OllamaMac, Windows, Linux
ollama run hf.co/unsloth/Llama-3_3-Nemotron-Super-49B-v1_5-GGUF:UD-IQ1_S - llama.cppMac, Windows, Linux
llama-server -hf unsloth/Llama-3_3-Nemotron-Super-49B-v1_5-GGUF:UD-IQ1_S - Docker Model RunnerMac, Windows, Linux
docker model run hf.co/unsloth/Llama-3_3-Nemotron-Super-49B-v1_5-GGUF:UD-IQ1_S
On a PC Graphics Card
| Video memory | Largest size that fits | Download | Memory at 8K |
|---|---|---|---|
| 6 GB | None worth running | needs 18 GB | |
| 8 GB | None worth running | needs 18 GB | |
| 10 GB | None worth running | needs 18 GB | |
| 11 GB | None worth running | needs 18 GB | |
| 12 GB | None worth running | needs 18 GB | |
| 16 GB | None worth running | needs 18 GB | |
| 20 GB | UD-IQ2_M | 17 GB | 18 GB |
| 24 GB | Q3_K_S | 22 GB | 23 GB |
| 32 GB | UD-Q4_K_XL | 30 GB | 31 GB |
| 48 GB | UD-Q6_K_XL | 43 GB | 44 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 | UD-IQ1_S | Too big | 27 GB | |
| M6 · 24 GB | UD-IQ2_XXS | Runs well | 16 GB | 4.6 |
| M6 · 32 GB | UD-IQ3_XXS | Runs well | 21 GB | 3.6 |
| M5 · 16 GB | UD-IQ1_S | Too big | 27 GB | |
| M5 · 24 GB | UD-IQ2_XXS | Runs well | 16 GB | 4.2 |
| M5 · 32 GB | UD-IQ3_XXS | Runs well | 21 GB | 3.3 |
| M5 Pro · 24 GB | UD-IQ2_XXS | Runs well | 16 GB | 8.3 |
| M5 Pro · 48 GB | UD-Q4_K_XL | Runs well | 31 GB | 4.7 |
| M5 Pro · 64 GB | Q4_K_S | Runs great | 31 GB | 4.9 |
| M5 Max · 36 GB | Q3_K_S | Runs well | 24 GB | 8.9 |
| M5 Max · 48 GB | UD-Q4_K_XL | Runs well | 31 GB | 9.1 |
| M5 Max · 64 GB | Q4_K_S | Runs great | 31 GB | 9.5 |
| M5 Max · 128 GB | Q8_0 | Runs great | 68 GB | 6.8 |
| M5 Ultra · 96 GB | UD-Q6_K_XL | Runs great | 45 GB | 13 |
| M5 Ultra · 256 GB | Q8_0 | Runs great | 68 GB | 13 |
| M5 Ultra · 512 GB | Q8_0 | Runs great | 68 GB | 13 |
| M4 · 16 GB | UD-IQ1_S | Too big | 27 GB | |
| M4 · 24 GB | UD-IQ2_XXS | Runs well | 16 GB | 3.3 |
| M4 · 32 GB | UD-IQ3_XXS | Runs well | 21 GB | 2.6 |
| M4 Pro · 24 GB | UD-IQ2_XXS | Runs well | 16 GB | 6.9 |
| M4 Pro · 48 GB | UD-Q4_K_XL | Runs well | 31 GB | 4.1 |
| M4 Pro · 64 GB | Q4_K_S | Runs great | 31 GB | 4.2 |
| M4 Max · 36 GB | Q3_K_S | Runs well | 24 GB | 7.4 |
| M4 Max · 48 GB | UD-Q4_K_XL | Runs well | 31 GB | 7.5 |
| M4 Max · 64 GB | Q4_K_S | Runs great | 31 GB | 7.9 |
| M4 Max · 96 GB | UD-Q6_K_XL | Runs great | 45 GB | 4.5 |
| M4 Max · 128 GB | Q8_0 | Runs great | 68 GB | 5.6 |
| M3 · 8 GB | UD-IQ1_S | Too big | 27 GB | |
| M3 · 16 GB | UD-IQ1_S | Too big | 27 GB | |
| M3 · 24 GB | UD-IQ2_XXS | Runs well | 16 GB | 2.8 |
| M3 Pro · 18 GB | UD-IQ1_M | Tight fit | 13 GB | 4.4 |
| M3 Pro · 36 GB | Q3_K_S | Runs well | 24 GB | 3 |
| M3 Max · 36 GB | Q3_K_S | Runs well | 24 GB | 5.7 |
| M3 Max · 48 GB | UD-Q4_K_XL | Runs well | 31 GB | 5.8 |
| M3 Max · 64 GB | Q4_K_S | Runs great | 31 GB | 6.1 |
| M3 Max · 96 GB | UD-Q6_K_XL | Runs great | 45 GB | 3.4 |
| M3 Max · 128 GB | Q8_0 | Runs great | 68 GB | 4.3 |
| M3 Ultra · 96 GB | UD-Q6_K_XL | Runs great | 45 GB | 8.3 |
| M3 Ultra · 256 GB | Q8_0 | Runs great | 68 GB | 8.1 |
| M3 Ultra · 512 GB | Q8_0 | Runs great | 68 GB | 8.1 |
| M2 · 8 GB | UD-IQ1_S | Too big | 27 GB | |
| M2 · 16 GB | UD-IQ1_S | Too big | 27 GB | |
| M2 · 24 GB | UD-IQ2_XXS | Runs well | 16 GB | 2.8 |
| M2 Pro · 16 GB | UD-IQ1_S | Too big | 27 GB | |
| M2 Pro · 32 GB | UD-IQ3_XXS | Runs well | 21 GB | 4.2 |
| M2 Max · 32 GB | UD-IQ3_XXS | Runs well | 21 GB | 7.9 |
| M2 Max · 64 GB | Q4_K_S | Runs great | 31 GB | 6.1 |
| M2 Max · 96 GB | UD-Q6_K_XL | Runs great | 45 GB | 4.4 |
| M2 Ultra · 64 GB | Q4_K_S | Runs great | 31 GB | 11 |
| M2 Ultra · 128 GB | Q8_0 | Runs great | 68 GB | 8.1 |
| M2 Ultra · 192 GB | Q8_0 | Runs great | 68 GB | 8.1 |
| M1 · 8 GB | UD-IQ1_S | Too big | 27 GB | |
| M1 · 16 GB | UD-IQ1_S | Too big | 27 GB | |
| M1 Pro · 16 GB | UD-IQ1_S | Too big | 27 GB | |
| M1 Pro · 32 GB | UD-IQ3_XXS | Runs well | 21 GB | 4 |
| M1 Max · 32 GB | UD-IQ3_XXS | Runs well | 21 GB | 7.3 |
| M1 Max · 64 GB | Q4_K_S | Runs great | 31 GB | 5.7 |
| M1 Ultra · 64 GB | Q4_K_S | Runs great | 31 GB | 9.9 |
| M1 Ultra · 128 GB | Q8_0 | Runs great | 68 GB | 7.5 |
Writes is words a second, estimated from the Mac's memory speed and the model's size; how, and how wrong.
Information
- Maker
- NVIDIA
- Released
- July 2025
- Parameters
- 50B
- Longest chat
- 128K tokens
- Reads pictures
- No
- Licence
- other
- Downloads this month
- 3,391
- Likes
- 11
- Sizes
- 26, 11 to 100 GB
- Kind
- Maker's release
Badge for the model card
Shows on Llama 3 3 Nemotron Super 49B V1 5'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.