Autor: Jakub Rusinowski · Ostatnia aktualizacja: 12 lipca 2026
Yes, comfortably — you'll have ~58 GB of headroom running Qwen 3.5 397B-A17B at IQ2_M (sub-Q4) (134 GB, ~48 tok/s (est.)).
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.
| VRAM | 192 GB unified memory |
| Memory Bandwidth | 800 GB/s |
| Quant | VRAM needed | Fits 192 GB? | Max context | Whole PC |
|---|---|---|---|---|
| F16 | 796.9 GB | ✗ No | — | — |
| Q8_0 | 424.7 GB | ✗ No | — | — |
| Q6_K | 328.5 GB | ✗ No | — | — |
| Q5_K_M | 284.3 GB | ✗ No | — | — |
| Q4_K_M | 242.6 GB | ✗ No | — | — |
| Q3_K_M | 172.1 GB | ✓ Yes | 32K | — |
| Q2_K | 133.4 GB | ✓ Yes | 64K | — |
Assumes a 4K-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 192 GB of unified memory. 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 Apple M2 Ultra: 8K 139 GB · 32K 151.7 GB · 128K 202.4 GB ✗. Past 128K it no longer fits 192 GB — a q8_0 KV cache buys roughly half of that back, which the calculator will price for you.
No compatible GPU? Qwen 3.5 on 8 GB of system RAM, CPU only: runs, but it is tight.
| Qwen 3.5 397B-A17B | IQ2_M (sub-Q4) · 134 GB · ~48 tok/s (est.) |
| Qwen 3.5 122B-A10B | Q4_K_M · 73.7 GB · ~50 tok/s (est.) |
| Qwen 3.5 35B-A3B | Q4_K_M · 21.1 GB · ~108 tok/s (est.) |
| Qwen 3.5 27B | Q4_K_M · 16.3 GB · ~24 tok/s (est.) |
| Qwen 3.5 9B | Q4_K_M · 5.4 GB · ~60 tok/s (est.) |
| Qwen 3.5 4B | Q4_K_M · 2.4 GB · ~105 tok/s (est.) |
| Qwen 3.5 2B | Q4_K_M · 1.2 GB · ~152 tok/s (est.) |
| Qwen 3.5 0.8B | Q4_K_M · 0.5 GB · ~212 tok/s (est.) |
At 2 hrs/day, buying (~$3,999) beats renting at $1.54/hr after about 3.6 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.
Yes, comfortably — you'll have ~58 GB of headroom running Qwen 3.5 397B-A17B at IQ2_M (sub-Q4) (134 GB, ~48 tok/s (est.)).
Qwen 3.5 397B-A17B at IQ2_M (sub-Q4) quantization (134 GB), estimated ~48 tokens/sec.
| Size | VRAM | Verdict | Speed |
|---|---|---|---|
| Qwen 3.5 397B-A17B | 134 GB | Runs | ~48 tok/s |
| Qwen 3.5 122B-A10B | 73.7 GB | Runs | ~50 tok/s |
| Qwen 3.5 35B-A3B | 21.1 GB | Runs | ~108 tok/s |
| Qwen 3.5 27B | 16.3 GB | Runs | ~24 tok/s |
| Qwen 3.5 9B | 5.4 GB | Runs | ~60 tok/s |
| Qwen 3.5 4B | 2.4 GB | Runs | ~105 tok/s |
| Qwen 3.5 2B | 1.2 GB | Runs | ~152 tok/s |
| Qwen 3.5 0.8B | 0.5 GB | Runs | ~212 tok/s |
← Can I Run It? | Qwen 3.5 Model Page | Apple M2 Ultra GPU Page | Check Your Hardware