Can I Run Qwen 3.5 397B-A17B on NVIDIA L40S?
Autor: Jakub Rusinowski · Ostatnia aktualizacja: 21 lipca 2026
No. Qwen 3.5 397B-A17B needs 139 GB at 8K context (134 GB weights + 5 GB KV cache/overhead) at IQ2_M (sub-Q4) and the NVIDIA L40S has only 48 GB. No quantization we publish closes a gap that size.
Get a personalized upgrade path →
The numbers
| Model | Qwen 3.5 397B-A17B |
| Parameters | 397B (17B active) |
| Quantization | IQ2_M (sub-Q4) |
| VRAM needed (8K context) | 139 GB |
| NVIDIA L40S VRAM | 48 GB |
| Verdict | Does not fit |
Qwen 3.5 397B-A17B on the NVIDIA L40S: VRAM by quantization
| Quant | VRAM needed | Fits 48 GB? | Max context | Whole PC |
|---|---|---|---|---|
| F16 | 799 GB | ✗ No | — | — |
| Q8_0 | 426.8 GB | ✗ No | — | — |
| Q6_K | 330.6 GB | ✗ No | — | — |
| Q5_K_M | 286.4 GB | ✗ No | — | — |
| Q4_K_M | 244.7 GB | ✗ No | — | — |
| Q3_K_M | 174.2 GB | ✗ No | — | — |
| Q2_K | 135.5 GB | ✗ No | — | — |
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 IQ2_M (sub-Q4) on the NVIDIA L40S: 8K 139 GB ✗ · 32K 151.7 GB ✗ · 128K 202.4 GB ✗. Past 8K it no longer fits 48 GB — a q8_0 KV cache buys roughly half of that back, which the calculator will price for you.
What to do next
Either step down to a smaller size in this family, or move to a Apple M2 Ultra (192 GB), the card with the least memory we track that holds it.
No compatible GPU? Qwen 3.5 on 8 GB of system RAM, CPU only: runs.
Cheapest card we track with enough VRAM for Qwen 3.5 397B-A17B:
Qwen 3.5 397B-A17B needs ~139 GB but this GPU has 48 GB. Rent a 2× A100 (160 GB)-class GPU by the hour instead of buying one:
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.
Other Qwen 3.5 sizes on the NVIDIA L40S
- Qwen 3.5 122B-A10B — runs at a lower quant (77.3 GB)
- Qwen 3.5 35B-A3B — runs (23.8 GB)
- Qwen 3.5 27B — runs (18.8 GB)
- Qwen 3.5 9B — runs (6.5 GB)
- Qwen 3.5 4B — runs (3.5 GB)
- Qwen 3.5 2B — runs (2.7 GB)
- Qwen 3.5 0.8B — runs (1.8 GB)
Qwen 3.5 397B-A17B on nearby hardware
- Apple M3 Pro (36 GB) — does not fit
- NVIDIA A100 80GB (PCIe) (80 GB) — does not fit
All Qwen 3.5 sizes on the NVIDIA L40S | Qwen 3.5 397B-A17B model page | NVIDIA L40S GPU page | VRAM calculator | Check your hardware