PC Build to Run GLM-4.7-Flash 30B-A3B Locally (Q5_K_M)
作者： Jakub Rusinowski · 最后更新： 2026年8月15日
A complete parts list — GPU, motherboard, RAM and PSU — worked backward from GLM-4.7-Flash 30B-A3B's VRAM requirement at Q5_K_M. Prices are 2026 street estimates and vary.
Requirements
| VRAM required | ~22 GB at Q5_K_M |
| System RAM | 32 GB |
| Estimated speed | ~128 tok/s |
| Complete build (est.) | $1,849 |
Weights plus framework overhead only — this model publishes no architecture we can read, so no KV cache is included. A real session needs more; the figure is a floor, not a target.
或在 Vast.ai 比较，低至 $0.35/小时 (typical low · varies)
作为亚马逊联盟成员，我们从符合条件的购买中获得收入。云 GPU 链接为推荐链接——我们可能获得佣金，您无需额外付费。
GPU: AMD Radeon RX 7900 XTX
AMD Radeon RX 7900 XTX's 24 GB covers GLM-4.7-Flash 30B-A3B at Q5_K_M (~22 GB needed), a tight fit — little headroom.
Motherboard: ASUS TUF Gaming B650-PLUS WIFI
B650 (AM5) — pairs with the AMD Ryzen 7 7700; DDR5, dual PCIe 4.0 x16 (x8/x8) for multi-GPU, 4 DIMM slots to 256 GB.
RAM: Corsair Vengeance 32GB (2x16GB) DDR5-6000 C30 (AMD EXPO)
32 GB DDR5-6000 (2x16GB) meets GLM-4.7-Flash 30B-A3B's 32 GB recommended system RAM.
PSU: Corsair RM650e (ATX 3.1)
650 W 80+ Gold covers AMD Radeon RX 7900 XTX (355 W) + CPU 65 W with 30% headroom (~546 W required).
Swap Alternatives
- GPU: AMD Radeon AI PRO R9700 — AMD Radeon AI PRO R9700's 32 GB covers GLM-4.7-Flash 30B-A3B at Q5_K_M (~22 GB needed), with headroom.
- GPU: NVIDIA GeForce RTX 5090 — NVIDIA GeForce RTX 5090's 32 GB covers GLM-4.7-Flash 30B-A3B at Q5_K_M (~22 GB needed), with headroom.
- RAM: Corsair Vengeance 64GB (2x32GB) DDR5-6000 C30 (AMD EXPO) — More headroom: 64 GB DDR5-6000.
- PSU: Corsair RM850e (ATX 3.1) — More headroom / future dual-GPU: 850 W 80+ Gold.
- The AMD Radeon RX 7900 XTX is a tight fit at Q5_K_M — consider an upgrade alternative for KV-cache/context headroom.
FAQ
What GPU do I need to run GLM-4.7-Flash 30B-A3B at Q5_K_M?
AMD Radeon RX 7900 XTX's 24 GB covers GLM-4.7-Flash 30B-A3B at Q5_K_M (~22 GB needed), a tight fit — little headroom.
How much VRAM does GLM-4.7-Flash 30B-A3B need at Q5_K_M?
About 22 GB of VRAM plus 32 GB of system RAM.
What power supply does a AMD Radeon RX 7900 XTX build need?
650 W 80+ Gold covers AMD Radeon RX 7900 XTX (355 W) + CPU 65 W with 30% headroom (~546 W required).
How much does a GLM-4.7-Flash 30B-A3B PC cost?
About $1,849 for a complete build (GPU, motherboard, RAM, PSU, CPU, case, SSD), at 2026 street-price estimates.
GLM-4.7-Flash 30B-A3B at Other Quant Levels
Other Popular Build Targets
- Qwen 3 32B PC build
- Gemma 4 31B PC build
- Qwen 3.6 35B-A3B PC build
- Qwen 3.5 35B-A3B PC build
- DeepSeek R1 Distill Qwen 32B PC build
Model & GPU
- GLM-4.7 / GLM-Z1 model page
- AMD Radeon RX 7900 XTX review
- Can I run GLM-4.7 / GLM-Z1 on AMD Radeon RX 7900 XTX?
More Ways to Build
← Budget-first build recommender | GLM-4.7 / GLM-Z1 model page | AMD Radeon RX 7900 XTX review