Alibaba Qwen · 9B total / 9B active · dense
Qwen3.5 9B needs about 7 GB at Q4_K_M. 17 of the 17 priced rigs we track can hold it. Here's the cheapest, the best value and the fastest. Prices & stock checked 2026-10-23. Prices change daily and some retailers block our checker, so always confirm on the store page.
15-core CPU / 16-core GPU, 24GB
Cheapest that fits. Lowest price we've seen on any tracked rig that holds Qwen3.5 9B at 4-bit (7 of 17 GB usable).
Suggested quant: Q8_0 (~28 tok/s)
Check price18-core CPU / 40-core GPU, 48GB
Best value. Lowest price per token/sec among complete systems where it fits comfortably: about $34 per tok/s at 4-bit.
Suggested quant: Q8_0 (~54 tok/s)
Check price32GB card
Fastest. Most memory bandwidth of the rigs that fit it (1,792 GB/s), so the quickest decode: ~214 tok/s at 4-bit.
Suggested quant: Q8_0 (~135 tok/s)
GPU price is card only: you still need a host PC.
Check priceUse Q4_K_M on most rigs: it needs ~7 GB. Q8_0 (~12 GB) is a bit more accurate and fits comfortably on 17 of these rigs, e.g. the Mac mini M5 Pro.
| Machine | Price | 4-bit fit & speed | Suggested quant | $ per tok/s |
|---|---|---|---|---|
| Mac mini M5 ProMac mini | $1,699 | Fits~48 tok/s | Q8_0 | $36 |
| GMKtec EVO-X2Strix Halo | $2,200 | Fits~37 tok/s | Q8_0 | $59 |
| Mac Studio M5 MaxMac Studio | $2,499 | Fits~70 tok/s | Q8_0 | $36 |
| Mac Studio M5 MaxMac Studio | $3,099 | Fits~90 tok/s | Q8_0 | $34 |
| Framework Desktop DIYStrix Halo | $3,449 | Fits~37 tok/s | Q8_0 | $92 |
| GMKtec EVO-X2Strix Halo | $3,500 | Fits~37 tok/s | Q8_0 | $94 |
| Minisforum MS-S1 MaxStrix Halo | $3,919 | Fits~37 tok/s | Q8_0 | $105 |
| Beelink GTR9 ProStrix Halo | $4,349 | Fits~37 tok/s | Q8_0 | $116 |
| Gigabyte GeForce RTX 4090 Windforce V2RTX 4090 (card only, needs a host PC) | $4,899 | Fits~133 tok/s | Q8_0 | $37 |
| Mac Studio M5 UltraMac Studio | $5,499 | Fits~139 tok/s | Q8_0 | $40 |
| Lenovo ThinkStation PGXDGX Spark / GB10 | $5,859 | Fits~39 tok/s | Q8_0 | $152 |
| ASUS Ascent GX10DGX Spark / GB10 | $5,999 | Fits~39 tok/s | Q8_0 | $156 |
| NVIDIA DGX SparkDGX Spark / GB10 | $6,950 | Fits~39 tok/s | Q8_0 | $180 |
| HP ZGX Nano G1nDGX Spark / GB10 | $7,699 | Fits~39 tok/s | Q8_0 | $200 |
| Gigabyte AI TOP ATOMDGX Spark / GB10 | $8,499 | Fits~39 tok/s | Q8_0 | $221 |
| NVIDIA GeForce RTX 5090 Founders EditionRTX 5090 (card only, needs a host PC) | $8,500 | Fits~214 tok/s | Q8_0 | $40 |
| PNY NVIDIA RTX PRO 6000 BlackwellRTX PRO 6000 (card only, needs a host PC) | $17,899 | Fits~214 tok/s | Q8_0 | $84 |
Sorted by lowest price. Speeds are estimated single-user decode at ~8K context. Full fit table, including rigs that don't fit: Qwen3.5 9B memory requirements. How we estimate
About 7 GB at Q4_K_M and 12 GB at Q8_0, including runtime overhead and ~8K tokens of context for one user. Longer context needs more.
Of the rigs we track, the Mac mini M5 Pro at $1,699, estimated ~48 tok/s at 4-bit. Prices checked 2026-10-23.
Q4_K_M is the default: about 7 GB. Q8_0 needs about 12 GB; use it only if you have the room and still get usable speed.
It has 9B active parameters per token, so speed is set by memory bandwidth: from ~37 tok/s on the slowest rig that fits to ~214 tok/s on the NVIDIA GeForce RTX 5090 Founders Edition. Estimates, not benchmarks.
One short email a week: what got cheaper, what's back in stock, and which new open models your machine can run. Unsubscribe anytime.
