Z.ai · 357B (32B active) · 専家混合（MoE）
Large GLM MoE with strong coding and 200K context 看看你的 GPU 或 Mac 跑不跑得動 GLM-4.6——最低 199.5 GB，建議 332.5 GB。
2025-09195K context
専家混合（MoE）
専家総数： 160
启用専家： 8
启用参数： 32.0B
| 量化 | 位元 | VRAM | 品質 | 状態 |
|---|---|---|---|---|
| Q2_K | 2 | 114.8 GB | low | — |
| Q3_K_M | 3 | 160.5 GB | moderate | — |
| Q4_K_M | 4 | 183.4 GB | good | — |
| Q5_K_M | 5 | 229.1 GB | good | — |
| Q6_K | 6 | 274.8 GB | excellent | — |
| Q8_0 | 8 | 366.2 GB | excellent | — |
| F16 | 16 | 732 GB | lossless | — |
Can I run GLM-4.6 locally?
- Can I run GLM-4.6 locally?
- GLM-4.6 needs about 199.5 GB of memory at a minimum and 332.5 GB recommended. Open this page to grade it against your GPU or Mac, then run it with runai, Ollama or LM Studio.
- How much VRAM does GLM-4.6 need?
- At Q4_K_M, GLM-4.6 uses about 183.4 GB of VRAM. Higher quants need more memory; lower quants fit tighter cards with a quality tradeoff.