Can GLM 4.5V run on the DGX Spark?
Yes. At 8K context the best fit is Q6_K, using about 84.4 GB of the 109 GB usable on the DGX Spark. Estimated speed ceiling: ~23 tokens/sec.
Memory by quant and context
| Quant | 4K ctx | 8K ctx | 32K ctx | Max tok/s (est., 8K) |
|---|---|---|---|---|
| 2-bit (Q2_K) | 39.0 GB Fits | 39.7 GB Fits | 44.1 GB Fits | 51 |
| 3-bit (Q3_K_M) | 50.4 GB Fits | 51.2 GB Fits | 55.5 GB Fits | 39 |
| 4-bit (Q4_K_M) • | 60.7 GB Fits | 61.4 GB Fits | 65.7 GB Fits | 32 |
| 5-bit (Q5_K_M) | 72.9 GB Fits | 73.6 GB Fits | 77.9 GB Fits | 27 |
| 6-bit (Q6_K) | 83.7 GB Fits | 84.4 GB Fits | 88.7 GB Fits | 23 |
| 8-bit (Q8_0) • | 107 GB Tight | 108 GB Tight | 112 GB No | 18 |
| 16-bit (F16/BF16) | 202 GB No | 203 GB No | 207 GB No | — |
• real GGUF file size from ggml-org/GLM-4.5V-GGUF. Speed is a bandwidth ceiling (273 GB/s), not a benchmark. Methodology.
More
GLM 4.5V on every GPU and Mac · Every model that runs on the DGX Spark · Other models that fit 64 GB VRAM · Try other settings in the calculator
Data: Hugging Face Hub (zai-org/GLM-4.5V), NVIDIA specs. Data as of 2026-11-17.