Can I Run IBM Granite 4.0 on NVIDIA L40S?
Written by Jakub Rusinowski · Last updated August 15, 2026
Yes, comfortably — Granite 4.0 Small-H 32B-A9B at Q4_K_M needs 22 GB at 8K context (19.3 GB weights + 2.7 GB KV cache/overhead), leaving ~26 GB of the NVIDIA L40S's 48 GB, at ~98 tok/s (est.), with room for up to 64K context.
See what else this hardware can run →
As an Amazon Associate we earn from qualifying purchases. Cloud GPU links are referral links — we may earn a commission at no extra cost to you.
NVIDIA L40S Specs
| VRAM | 48 GB |
| Memory Bandwidth | 864 GB/s |
Granite 4.0 Small-H 32B-A9B on the NVIDIA L40S: VRAM by quantization
| Quant | VRAM needed | Fits 48 GB? | Max context | Whole PC |
|---|---|---|---|---|
| F16 | 66.6 GB | ✗ No | — | — |
| Q8_0 | 36.6 GB | ✓ Yes | 32K | — |
| Q6_K | 28.9 GB | ✓ Yes | 64K | — |
| Q5_K_M | 25.3 GB | ✓ Yes | 64K | — |
| Q4_K_M | 22 GB | ✓ Yes | 64K | — |
| Q3_K_M | 16.3 GB | ✓ Yes | 128K | — |
| Q2_K | 13.2 GB | ✓ Yes | 128K | — |
Assumes an 8K-token context with an f16 KV cache, on an estimated architecture — this model publishes no config we can read. A longer window needs more; a quantized KV cache needs less. “Max context” is the largest window that still fits in 48 GB. Figures are estimates from parameter count, quantization and memory bandwidth — the analyzer lets you tune KV-cache quant and context.
Context costs VRAM too. At Q4_K_M on the NVIDIA L40S: 8K 22 GB · 32K 27.5 GB · 128K 49.6 GB ✗. Past 128K it no longer fits 48 GB — a q8_0 KV cache buys roughly half of that back, which the calculator will price for you.
No compatible GPU? Granite 4.0 Small-H 32B-A9B on 24 GB of system RAM, CPU only: runs.
IBM Granite 4.0 Sizes That Fit the NVIDIA L40S (at 8K context)
| Granite 4.0 Small-H 32B-A9B | Q4_K_M · 22 GB at 8K context · ~98 tok/s (est.) |
At 2 hrs/day, buying (~$7,499) beats renting at $0.34/hr after about 30.6 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.
FAQ
Can I run IBM Granite 4.0 on the NVIDIA L40S?
Yes, comfortably — Granite 4.0 Small-H 32B-A9B at Q4_K_M needs 22 GB at 8K context (19.3 GB weights + 2.7 GB KV cache/overhead), leaving ~26 GB of the NVIDIA L40S's 48 GB, at ~98 tok/s (est.), with room for up to 64K context.
What's the best IBM Granite 4.0 size for the NVIDIA L40S?
Granite 4.0 Small-H 32B-A9B at Q4_K_M quantization needs 22 GB at 8K context (19.3 GB weights + 2.7 GB KV cache/overhead), estimated ~98 tokens/sec, up to 64K context.
IBM Granite 4.0 on Other GPUs
- IBM Granite 4.0 on NVIDIA RTX 6000 Ada Generation
- IBM Granite 4.0 on Apple M3 Pro
- IBM Granite 4.0 on NVIDIA GeForce RTX 5090
- IBM Granite 4.0 on Apple M4
- IBM Granite 4.0 on Apple M2 Pro
Popular Models on the NVIDIA L40S
VRAM Tier
Troubleshooting
Buying Guide
← Can I Run It? | IBM Granite 4.0 Model Page | NVIDIA L40S GPU Page | Check Your Hardware