zai org GLM 4.5 needs about 222.1 GB of memory at Q4_K_M (a 218.5 GB file) with an 8K context window, which is more than a 48 GB GPU holds. This repository has 26 quants, from 0.7 GB to 381.0 GB (Q8_0). Longer context needs more: at 32K tokens the KV cache alone adds about 12.5 GB.
GLM‑4.5, a 358‑billion‑parameter mixture‑of‑experts model
GLM‑4.5 is a foundation model with 358 billion total parameters and a trained context of 131 072 tokens. It is released under the MIT license and supports hybrid reasoning for intelligent agents. The model is available in multiple GGUF quantizations.
Good at
- Large‑context reasoning
- Hybrid thinking and non‑thinking modes
- Agent‑oriented tasks
Limits
- Requires large VRAM for higher quants
- No vision support
- Limited to 131 072 token context
Best for
- Complex reasoning workflows
- Tool‑enabled agent applications
- High‑capacity text generation
Who it’s for. Developers building intelligent agents or large‑context applications who need a 358 billion‑parameter model. The model is suitable for commercial use under the MIT license.
Quants & memory
| Quant | File | Est. memory | vs VRAM | Result | |
|---|---|---|---|---|---|
| — | 0.7 GB | 4.3 GB | Fits | Get file → | |
| IQ1_S | 77.1 GB (2 parts) | 80.7 GB | Too big | Get file → | |
| IQ1_M | 80.4 GB (3 parts) | 84.0 GB | Too big | Get file → | |
| IQ2_XXS | 89.1 GB (3 parts) | 92.7 GB | Too big | Get file → | |
| IQ2_XS | 101.8 GB (3 parts) | 105.5 GB | Too big | Get file → | |
| IQ2_S | 102.7 GB (3 parts) | 106.3 GB | Too big | Get file → | |
| IQ2_M | 115.9 GB (3 parts) | 119.5 GB | Too big | Get file → | |
| Q2_K | 127.8 GB (4 parts) | 131.4 GB | Too big | Get file → | |
| Q2_K_L | 128.5 GB (4 parts) | 132.1 GB | Too big | Get file → | |
| IQ3_XXS | 143.3 GB (4 parts) | 146.9 GB | Too big | Get file → | |
| IQ3_XS | 148.9 GB (4 parts) | 152.5 GB | Too big | Get file → | |
| Q3_K_S | 157.3 GB (4 parts) | 160.9 GB | Too big | Get file → | |
| IQ3_M | 164.9 GB (5 parts) | 168.6 GB | Too big | Get file → | |
| Q3_K_M | 165.0 GB (5 parts) | 168.6 GB | Too big | Get file → | |
| Q3_K_L | 170.6 GB (5 parts) | 174.2 GB | Too big | Get file → | |
| Q3_K_XL | 171.3 GB (5 parts) | 174.9 GB | Too big | Get file → | |
| IQ4_XS | 192.8 GB (5 parts) | 196.5 GB | Too big | Get file → | |
| IQ4_NL | 203.7 GB (6 parts) | 207.4 GB | Too big | Get file → | |
| Q4_0 | 206.5 GB (6 parts) | 210.1 GB | Too big | Get file → | |
| Q4_K_S | 210.8 GB (6 parts) | 214.4 GB | Too big | Get file → | |
| Q4_K_M | 218.5 GB (6 parts) | 222.1 GB | Too big | Get file → | |
| Q4_1 | 225.1 GB (6 parts) | 228.7 GB | Too big | Get file → | |
| Q5_K_S | 247.4 GB (7 parts) | 251.1 GB | Too big | Get file → | |
| Q5_K_M | 255.3 GB (7 parts) | 258.9 GB | Too big | Get file → | |
| Q6_K | 294.8 GB (8 parts) | 298.5 GB | Too big | Get file → | |
| Q8_0 | 381.0 GB (10 parts) | 384.6 GB | Too big | Get file → |
For 8 GB VRAM, use the 0.69 GB quant. For 16 GB VRAM, use IQ1_S (77.11 GB) or IQ1_M (80.37 GB). For 32 GB VRAM, Q4_0 (206.49 GB) or Q4_K_S (210.78 GB) are recommended. Estimates are approximate.
What GPU do you need for zai org GLM 4.5?
These figures add the file size, an fp16 KV cache at an 8K context window (worked out from this model’s own header) and about 0.5 GB for the runtime. They are estimates, so leave a little headroom, and expect to use less context or a smaller quant if you are right at the limit.
- 4 GB VRAM: no quant fits at an 8K context.
- 6 GB VRAM: the best quant that fits is the main file (0.7 GB file, about 4.3 GB in use).
- 8 GB VRAM: the best quant that fits is the main file (0.7 GB file, about 4.3 GB in use).
- 12 GB VRAM: the best quant that fits is the main file (0.7 GB file, about 4.3 GB in use).
- 16 GB VRAM: the best quant that fits is the main file (0.7 GB file, about 4.3 GB in use).
- 24 GB VRAM: the best quant that fits is the main file (0.7 GB file, about 4.3 GB in use).
- 32 GB VRAM: the best quant that fits is the main file (0.7 GB file, about 4.3 GB in use).
- 48 GB VRAM: the best quant that fits is the main file (0.7 GB file, about 4.3 GB in use).
Which zai org GLM 4.5 quant should you choose?
Q4_K_M is the usual starting point for zai org GLM 4.5: it balances file size and quality well. Here it is a 218.5 GB file.
The Q8_0 file is 381.0 GB, about 1.7 times the size of Q4_K_M. It keeps the weights closer to the original, which only matters if you have the memory to spare.
If memory is tight, the smallest file is the first one in the table at 0.7 GB, with the quality trade-offs that come with fewer bits.
New to quant names? Read GGUF quantization explained.
Specifications
Architecture from the GGUF header
Source
Frequently asked questions about zai org GLM 4.5
How much VRAM does zai org GLM 4.5 need?
About 222.1 GB at Q4_K_M with an 8K context window. That is the 218.5 GB file, plus the KV cache and roughly 0.5 GB of runtime overhead. Smaller quants need less and longer context needs more.
Can I run zai org GLM 4.5 on an 8 GB GPU?
Yes, with the smaller files: the 0.7 GB file needs about 4.3 GB at 8K context. Larger quants in this repository do not fit in 8 GB.
What is the file size of zai org GLM 4.5 GGUF?
It depends on the quant: file is 0.7 GB, IQ1_S is 77.1 GB, IQ1_M is 80.4 GB, IQ2_XXS is 89.1 GB, IQ2_XS is 101.8 GB, IQ2_S is 102.7 GB and 20 more.
What context length does zai org GLM 4.5 support?
The GGUF header lists a trained context length of 131,072 tokens (128K). The KV cache grows with the context you use: about 3.1 GB at 8K and 12.5 GB at 32K.
Where can I download zai org GLM 4.5 GGUF?
The files are on Hugging Face in the bartowski/zai-org_GLM-4.5-GGUF repository. The download buttons on this page open the file on Hugging Face. ggufmodels.com does not host any model files.
What license does zai org GLM 4.5 use?
Hugging Face lists the license as mit. Licenses can differ between the quantized files and the original model, so check the repository and the base model before commercial use.
Can I run GLM‑4.5 on 8 GB VRAM?
Only the 0.69 GB quant can be loaded into 8 GB VRAM. Larger quants will exceed the memory limit.
Does GLM‑4.5 support vision?
No, the model has no vision capabilities.
What license does GLM‑4.5 use?
The model is released under the MIT open‑source license.