unsloth
DeepSeek V3.1
DeepSeek V3.1 is a 671.0B mixture-of-experts deepseek2 model from unsloth, with about 121.6B parameters active per token. Configured context length is 163,840 tokens. License: mit. Quant sizes are the GGUF file bytes published on Hugging Face when a file exists.
Parameters671B / 122B active
Context160K
Attentionmla
Licensemit
61 layers · 128 KV heads · head dim 56 · GGUF repo · config
Quant file sizes
| Quant | File size | Source |
|---|---|---|
| BF16 | 1250.1 GB | Hugging Face file bytes |
| UD-Q8_K_XL | 727.0 GB | Hugging Face file bytes |
| Q8_0 | 664.3 GB | Hugging Face file bytes |
| UD-Q6_K_XL | 535.0 GB | Hugging Face file bytes |
| Q6_K | 513.5 GB | Hugging Face file bytes |
| UD-Q5_K_XL | 451.3 GB | Hugging Face file bytes |
| Q5_K_M | 443.5 GB | Hugging Face file bytes |
| Q5_K_S | 430.9 GB | Hugging Face file bytes |
| UD-Q4_K_XL | 360.3 GB | Hugging Face file bytes |
| Q4_K_M | 377.6 GB | Hugging Face file bytes |
| Q4_K_S | 354.9 GB | Hugging Face file bytes |
| IQ4_NL | 353.1 GB | Hugging Face file bytes |
| IQ4_XS | 333.7 GB | Hugging Face file bytes |
| Q4_1 | 391.9 GB | Hugging Face file bytes |
| Q4_0 | 354.0 GB | Hugging Face file bytes |
| UD-Q3_K_XL | 279.4 GB | Hugging Face file bytes |
| Q3_K_M | 298.4 GB | Hugging Face file bytes |
| Q3_K_S | 270.5 GB | Hugging Face file bytes |
| UD-IQ3_XXS | 260.5 GB | Hugging Face file bytes |
| UD-Q2_K_XL | 238.2 GB | Hugging Face file bytes |
| Q2_K | 457.8 GB | Hugging Face file bytes |
| UD-IQ2_XXS | 210.4 GB | Hugging Face file bytes |
| UD-IQ1_M | 192.6 GB | Hugging Face file bytes |
| UD-IQ1_S | 179.1 GB | Hugging Face file bytes |
On common machines
| Device | Verdict | Quant | Speed |
|---|---|---|---|
| GeForce RTX 4060 Ti 8GB | Won't fit | UD-IQ1_S | — |
| GeForce RTX 3060 12GB | Won't fit | UD-IQ1_S | — |
| GeForce RTX 4060 Ti 16GB | Won't fit | UD-IQ1_S | — |
| GeForce RTX 4090 | Won't fit | UD-IQ1_S | — |
| GeForce RTX 5090 | Won't fit | UD-IQ1_S | — |
| Apple M4 Pro | Won't fit | UD-IQ1_S | — |
| Apple M5 | Won't fit | UD-IQ1_S | — |
| CPU only · DDR5-5600 | Won't fit | UD-IQ1_S | — |
Smallest devices that fit it fully
No GPU or Mac in the catalog fits this model inside the default memory budget. A partial offload or a larger machine is required.
Run it
Commands for the GeForce RTX 4060 Ti 16GB estimate (UD-IQ1_S).
llama.cpp
llama-server -hf unsloth/DeepSeek-V3.1-GGUF:UD-IQ1_S -c 2048 -ngl 0
Ollama
ollama run deepseek-v3.1
LM Studio
lms get unsloth/DeepSeek-V3.1-GGUF # Then load quant UD-IQ1_S