GLM-5.3
Z.ai's flagship open model for agentic and coding work, with a 1M token context. Large, so plan a dedicated multi-GPU server.
Overview
GLM-5.3 is a mixture-of-experts model with 256 experts and 8 active per token, about 753B parameters in total. It targets agentic coding and tool use and supports a context window of 1,048,576 tokens.
Hardware fit
The smallest practical builds start around 217 GB (UD-IQ1_S) and UD-Q4_K_XL is 467 GB, so this needs a full multi-GPU node.
| Quantization | Weights |
|---|---|
| UD-IQ1_S | 216.7 GB |
| UD-Q2_K_XL | 253.9 GB |
| UD-Q4_K_XL | 467.3 GB |
Sizes are the GGUF file sizes from unsloth/GLM-5.3-GGUF. Budget extra VRAM for the KV cache, which grows with context size and parallel requests.
Deploying on Loadcore
- In the console, open Models and search for
unsloth/GLM-5.3-GGUF. - Pick the quantization that fits the target server and start the download.
- Start an instance with the context size you need and GPU layers set to offload the full model.
- Point your tools at
http://your-server:9090/v1with your personal key.
Recommended settings
| Parameter | Value |
|---|---|
| temperature | 1.0 |
| top_p | 0.95 |
| top_k | 0 |
| repeat_penalty | 1.0 |
| min_p | 0.0 |
The model card gives no single recommended setting. It reports most evaluations with temperature 1.0 and top_p 0.95, which are the values below. Review the GLM-5.3 license terms before use.
Source: zai-org/GLM-5.3.