作者: Jakub Rusinowski · 最后更新: 2026年4月29日
Yes, comfortably — you'll have ~6 GB of headroom running Granite 4.1 30B at FP8 (Q4 est. ~12-15GB, unconfirmed) (18 GB, ~52 tok/s (est.)).
在亚马逊查看价格 — NVIDIA GeForce RTX 3090 24GB
| VRAM | 24 GB |
| Memory Bandwidth | 936 GB/s |
| Quant | VRAM needed | Fits 24 GB? | Max context |
|---|---|---|---|
| F16 | 70.4 GB | ✗ No | — |
| Q8_0 | 42.3 GB | ✗ No | — |
| Q6_K | 35 GB | ✗ No | — |
| Q5_K_M | 31.7 GB | ✗ No | — |
| Q4_K_M | 28.5 GB | ✗ No | — |
| Q3_K_M | 23.2 GB | ✓ Yes | 4K |
| Q2_K | 20.3 GB | ✓ Yes | 4K |
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.
| Granite 4.1 30B | FP8 (Q4 est. ~12-15GB, unconfirmed) · 18 GB · ~52 tok/s (est.) |
| Granite 4.1 8B | Q4_K_M · 5 GB · ~187 tok/s (est.) |
| Granite 4.1 3B | Q4_K_M · 2 GB · ~400 tok/s (est.) |
At 2 hrs/day, buying (~$1,499) beats renting at $0.34/hr after about 6.1 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 ~6 GB of headroom running Granite 4.1 30B at FP8 (Q4 est. ~12-15GB, unconfirmed) (18 GB, ~52 tok/s (est.)).
Granite 4.1 30B at FP8 (Q4 est. ~12-15GB, unconfirmed) quantization (18 GB), estimated ~52 tokens/sec.
← Can I Run It? | IBM Granite 4.1 Model Page | NVIDIA GeForce RTX 3090 GPU Page | Check Your Hardware