Loading…
Loading…
Models / GLM-5.3-Flash
1 model (~321B) · ↓ 6.7M downloads a month · from $2.36/h on 4x Nvidia A40 right now. Each model's versions do the same job; quantized ones (AWQ, FP8, 4-bit) need less GPU memory and run on cheaper machines.
| Version | Runtime | Weights | GPU memory | Cheapest machine now |
|---|---|---|---|---|
| Originalzai-org/GLM-5.3-Flash | vLLM | 328 GB | 373.8 GB | $6.47/h 4x RTX PRO 6000 |
| NVFP4 (nvidia)nvidia/GLM-5.3-Flash-NVFP4 | vLLM | 204 GB | 236.1 GB | $4.73/h 8x RTX 5090 |
| GLM5.3 Flash E224 DGX Spark (autotrust)autotrust/GLM5.3-Flash-E224-DGX-Spark | vLLM | 152 GB | 177.3 GB | $2.36/h 4x Nvidia A40 |
| CYBERSECURITY W4A16 (dealignai)dealignai/GLM-5.3-Flash-CYBERSECURITY-W4A16 | vLLM | 195 GB | 225.2 GB | $4.73/h 8x RTX 5090 |
Model details come from Hugging Face and are refreshed daily; prices follow the live offers. GPU memory is GridPup's estimate: the weights, room for an 8,192-token conversation (chat models) and the runtime's own memory. Large models are split across the GPUs of one machine ("4x RTX 4090").