unsloth
GLM 4.6
GLM 4.6 is a 356.8B mixture-of-experts glm4moe model from unsloth, with about 71.4B parameters active per token. Configured context length is 202,752 tokens. License: mit. Quant sizes are the GGUF file bytes published on Hugging Face when a file exists.
Parameters357B / 71.4B active
Context198K
Attentiongqa
Licensemit
92 layers · 8 KV heads · head dim 128 · GGUF repo · config
Quant file sizes
| Quant | File size | Source |
|---|---|---|
| BF16 | 664.7 GB | Hugging Face file bytes |
| UD-Q8_K_XL | 363.0 GB | Hugging Face file bytes |
| Q8_0 | 353.3 GB | Hugging Face file bytes |
| UD-Q6_K_XL | 278.4 GB | Hugging Face file bytes |
| Q6_K | 273.0 GB | Hugging Face file bytes |
| UD-Q5_K_XL | 234.7 GB | Hugging Face file bytes |
| Q5_K_M | 235.8 GB | Hugging Face file bytes |
| Q5_K_S | 229.0 GB | Hugging Face file bytes |
| UD-Q4_K_XL | 189.7 GB | Hugging Face file bytes |
| Q4_K_M | 200.8 GB | Hugging Face file bytes |
| Q4_K_S | 188.9 GB | Hugging Face file bytes |
| IQ4_NL | 187.7 GB | Hugging Face file bytes |
| IQ4_XS | 177.6 GB | Hugging Face file bytes |
| Q4_1 | 208.3 GB | Hugging Face file bytes |
| Q4_0 | 188.3 GB | Hugging Face file bytes |
| UD-Q3_K_XL | 147.4 GB | Hugging Face file bytes |
| Q3_K_M | 158.9 GB | Hugging Face file bytes |
| Q3_K_S | 143.8 GB | Hugging Face file bytes |
| UD-IQ3_XXS | 134.8 GB | Hugging Face file bytes |
| UD-Q2_K_XL | 125.4 GB | Hugging Face file bytes |
| Q2_K | 243.5 GB | Hugging Face file bytes |
| UD-IQ2_XXS | 107.5 GB | Hugging Face file bytes |
| UD-IQ1_M | 100.0 GB | Hugging Face file bytes |
| UD-IQ1_S | 90.3 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
- NVIDIA H200 SXM — Tight fit, UD-Q2_K_XL, 88–120 tok/s
- NVIDIA B200 — Runs, UD-Q3_K_XL, 127–173 tok/s
Run it
Commands for the GeForce RTX 4060 Ti 16GB estimate (UD-IQ1_S).
llama.cpp
llama-server -hf unsloth/GLM-4.6-GGUF:UD-IQ1_S -c 2048 -ngl 0
No matching name in the Ollama library snapshot.
LM Studio
lms get unsloth/GLM-4.6-GGUF # Then load quant UD-IQ1_S