GLM 5 needs about 457.8 GB of memory at Q4_K_M (a 455.9 GB file) with an 8K context window, which is more than a 48 GB GPU holds. This repository has 28 quants, from 176.1 GB (TQ1_0) to 1508.0 GB (BF16). Longer context needs more: at 32K tokens the KV cache alone adds about 5.6 GB.
GLM‑5: 744 B parameters, 202 k context, MIT licensed
GLM‑5 is a 744 B parameter model with 202 752 token context. It uses glm‑dsa architecture and DeepSeek Sparse Attention. The model is released under the MIT license and supports large‑context reasoning and agentic tasks.
Good at
- Long‑context reasoning
- Agentic task planning
- Complex systems engineering
Limits
- Requires large VRAM for full precision
- No vision support
- License is MIT (open source)
Best for
- Research on long‑horizon tasks
- Fine‑tuning for specialized agents
- Benchmarking large‑context models
Who it’s for. Researchers and developers needing a large‑context LLM for complex reasoning. The model is open source and can be fine‑tuned for specialized applications.
Quants & memory
| Quant | File | Est. memory | vs VRAM | Result | |
|---|---|---|---|---|---|
| TQ1_0 | 176.1 GB | 178.1 GB | Too big | Get file → | |
| IQ1_S | 203.7 GB (6 parts) | 205.6 GB | Too big | Get file → | |
| IQ1_M | 223.9 GB (6 parts) | 225.8 GB | Too big | Get file → | |
| IQ2_XXS | 241.1 GB (6 parts) | 243.0 GB | Too big | Get file → | |
| IQ2_M | 254.8 GB (7 parts) | 256.7 GB | Too big | Get file → | |
| Q2_K | 276.0 GB (7 parts) | 277.9 GB | Too big | Get file → | |
| Q2_K_L | 276.2 GB (7 parts) | 278.1 GB | Too big | Get file → | |
| Q2_K_XL | 281.4 GB (7 parts) | 283.3 GB | Too big | Get file → | |
| IQ3_XXS | 304.5 GB (8 parts) | 306.4 GB | Too big | Get file → | |
| Q3_K_S | 326.3 GB (8 parts) | 328.2 GB | Too big | Get file → | |
| Q3_K_XL | 332.5 GB (8 parts) | 334.4 GB | Too big | Get file → | |
| Q3_K_M | 360.3 GB (9 parts) | 362.2 GB | Too big | Get file → | |
| IQ4_XS | 402.9 GB (10 parts) | 404.8 GB | Too big | Get file → | |
| MXFP4_MOE | 410.6 GB (10 parts) | 412.5 GB | Too big | Get file → | |
| IQ4_NL | 425.9 GB (10 parts) | 427.8 GB | Too big | Get file → | |
| Q4_0 | 427.1 GB (10 parts) | 429.1 GB | Too big | Get file → | |
| Q4_K_S | 428.4 GB (10 parts) | 430.3 GB | Too big | Get file → | |
| Q4_K_XL | 430.7 GB (10 parts) | 432.6 GB | Too big | Get file → | |
| Q4_K_M | 455.9 GB (11 parts) | 457.8 GB | Too big | Get file → | |
| Q4_1 | 472.7 GB (11 parts) | 474.6 GB | Too big | Get file → | |
| Q5_K_S | 519.7 GB (12 parts) | 521.6 GB | Too big | Get file → | |
| Q5_K_M | 535.2 GB (12 parts) | 537.1 GB | Too big | Get file → | |
| Q5_K_XL | 536.2 GB (12 parts) | 538.1 GB | Too big | Get file → | |
| Q6_K | 619.4 GB (14 parts) | 621.3 GB | Too big | Get file → | |
| Q6_K_XL | 645.2 GB (15 parts) | 647.1 GB | Too big | Get file → | |
| Q8_0 | 801.3 GB (18 parts) | 803.3 GB | Too big | Get file → | |
| Q8_K_XL | 868.8 GB (19 parts) | 870.7 GB | Too big | Get file → | |
| BF16 | 1508.0 GB (33 parts) | 1509.9 GB | Too big | Get file → |
For 8 GB VRAM, use Q4_0 (≈429 GB memory at 8k context). For 16 GB VRAM, Q4_K_S (≈430 GB). For 24 GB VRAM, Q4_K_M (≈458 GB). Larger quants like Q6_K or Q8_0 require more than 32 GB VRAM.
What GPU do you need for GLM 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: no quant fits at an 8K context.
- 8 GB VRAM: no quant fits at an 8K context.
- 12 GB VRAM: no quant fits at an 8K context.
- 16 GB VRAM: no quant fits at an 8K context.
- 24 GB VRAM: no quant fits at an 8K context.
- 32 GB VRAM: no quant fits at an 8K context.
- 48 GB VRAM: no quant fits at an 8K context.
Which GLM 5 quant should you choose?
Q4_K_M is the usual starting point for GLM 5: it balances file size and quality well. Here it is a 455.9 GB file.
The Q8_0 file is 801.3 GB, about 1.8 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 TQ1_0 at 176.1 GB, with the quality trade-offs that come with fewer bits. Lower bit counts mean smaller files and more quality loss.
New to quant names? Read GGUF quantization explained.
Specifications
Architecture from the GGUF header
Source
Frequently asked questions about GLM 5
How much VRAM does GLM 5 need?
About 457.8 GB at Q4_K_M with an 8K context window. That is the 455.9 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 GLM 5 on an 8 GB GPU?
Not at an 8K context. The smallest file needs about 178.1 GB, so you would need a larger GPU or CPU offloading.
What is the file size of GLM 5 GGUF?
It depends on the quant: TQ1_0 is 176.1 GB, IQ1_S is 203.7 GB, IQ1_M is 223.9 GB, IQ2_XXS is 241.1 GB, IQ2_M is 254.8 GB, Q2_K is 276.0 GB and 22 more.
What context length does GLM 5 support?
The GGUF header lists a trained context length of 202,752 tokens (198K). The KV cache grows with the context you use: about 1.4 GB at 8K and 5.6 GB at 32K.
Where can I download GLM 5 GGUF?
The files are on Hugging Face in the unsloth/GLM-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 GLM 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‑5 on 8 GB VRAM?
Use the Q4_0 quant, which requires about 429 GB of VRAM at 8k context. This is the smallest quant available for 8 GB VRAM.
Does GLM‑5 support vision tasks?
No, GLM‑5 does not include vision capabilities.