作者： Jakub Rusinowski · 最后更新： 2024年11月12日
Yes, comfortably — you'll have ~4.5 GB of headroom running Qwen 2.5 Coder 32B at Q4_K_M (19.5 GB, ~52 tok/s (est.)) with room for up to 8K context.
在亚马逊查看价格 — NVIDIA GeForce RTX 4090 24GB
| VRAM | 24 GB |
| Memory Bandwidth | 1008 GB/s |
| Quant | VRAM needed | Fits 24 GB? | Max context |
|---|---|---|---|
| F16 | 65.9 GB | ✗ No | — |
| Q8_0 | 35.9 GB | ✗ No | — |
| Q6_K | 28.1 GB | ✗ No | — |
| Q5_K_M | 24.6 GB | ✗ No | — |
| Q4_K_M | 21.2 GB | ✓ Yes | 8K |
| Q3_K_M | 15.5 GB | ✓ Yes | 32K |
| Q2_K | 12.4 GB | ✓ Yes | 32K |
VRAM needed assumes a 4K-token context with an f16 KV cache; “Max context” is the largest window that still fits in 24 GB. Figures are estimates from parameter count, quantization and memory bandwidth — the analyzer lets you tune KV-cache quant and context.
| Qwen 2.5 Coder 32B | Q4_K_M · 19.5 GB · ~52 tok/s (est.) |
| Qwen 2.5 14B Instruct | Q4_K_M · 9.5 GB · ~106 tok/s (est.) |
| Qwen 2.5 7B Instruct | Q4_K_M · 4.8 GB · ~210 tok/s (est.) |
At 2 hrs/day, buying (~$1,599) beats renting at $0.34/hr after about 6.5 years.
Affiliate links — we may earn a commission if you sign up, at no extra cost to you.
Cloud rates verified 2026-07 — estimates, and marketplace prices vary. Buying price is GPU MSRP only, not a full PC.
Yes, comfortably — you'll have ~4.5 GB of headroom running Qwen 2.5 Coder 32B at Q4_K_M (19.5 GB, ~52 tok/s (est.)) with room for up to 8K context.
Qwen 2.5 Coder 32B at Q4_K_M quantization (19.5 GB), estimated ~52 tokens/sec, up to 8K context.
← Can I Run It? | Qwen 2.5 Family Model Page | NVIDIA GeForce RTX 4090 GPU Page | Check Your Hardware