Run Qwen2.5-Coder 32B on a VPS.
Prices checked
The cheapest VPS that runs Qwen2.5-Coder 32B is Leaseweb VPS 4 at $26.13 a month, with 48 GB of RAM for the 23.4 GB it needs, checked .
Best plans for Qwen2.5-Coder 32B
Most tokens per second for your money, comfortable fits first.
- 1
VPS 4 · 12 vCPU / 48 GB · Europe
Leaseweb · 12C · 48 GB · ~1.2–2.4 tok/s$26.13/mo - 2
VPS 4 · 12 vCPU / 48 GB · US
Leaseweb · 12C · 48 GB · ~1.2–2.4 tok/s$31.49/mo - 3
VPS 4 · 12 vCPU / 48 GB · UK
Leaseweb · 12C · 48 GB · ~1.2–2.4 tok/s$32.55/mo - 4$32.95/mo
- 5
VPS 5 · 16 vCPU / 64 GB · Europe
Leaseweb · 16C · 64 GB · ~1.2–2.4 tok/s$34.18/mo - 6
VPS 4 · 12 vCPU / 48 GB · Singapore
Leaseweb · 12C · 48 GB · ~1.2–2.4 tok/s$39.95/mo
Hetzner · 16C · 32 GB · ~1.2–2.4 tok/sWhat it needs
RAM needed
23.4 GB
CPU inference at Q4_K_M
Weights
19.9 GB
Q4_K_M
KV cache
2 GB
8K context
OS and runtime
1.5 GB
headroom
Open coding model that traded blows with GPT-4o at launch; needs 32 GB.
- Size
- 33B parameters
- Context
- 128K tokens
- Kind
- code
- Released
- 2024-11
- License
- Apache-2.0
Facts from Hugging Face: exact GGUF file sizes, KV geometry from the GGUF header, 553,084 downloads.
Or rent it as an API
every price for Qwen2.5 Coder 32B Instruct14 providers sell Qwen2.5 Coder 32B Instruct.
Alibaba Cloud's own price
$0.75per M tokens
Cheapest price
$0.10per M tokens
Nscale
What the VPS price buys
275M tokens a month
at $26.13 a month
Or rent a GPU
every GPU price1 provider rents the RTX A5000, with 24 GB of GPU memory against the 23.4 GB Qwen2.5-Coder 32B needs.
Cheapest board that fits
RTX A5000
24 GB of GPU memory
On demand
$0.16per GPU an hour
RunPod
Held for a month
$116.80
730 hours of RTX A5000 · Community Cloud
At 5.3 million tokens a month, what this box generates in 730 hours at its estimated rate, the API costs you $0.50, this VPS $26.13 and the GPU $116.80. Above about 275 million tokens a month the VPS costs less than the API, which is more than this box generates in a month.
Or buy hardware
all machines8 reference machines fit Qwen2.5-Coder 32B.
Mac mini (M6, 32 GB)
Tight fit32 GB · 153 GB/s · ~4–6.2 tok/s
$1,299= 50 mo of VPSFramework Desktop (Ryzen AI Max+ 395, 64 GB)
Approximate64 GB · 256 GB/s · ~6.8–10 tok/s
$1,959= 75 mo of VPSMac mini (M5 Pro, 64 GB)
64 GB · 307 GB/s · ~8.1–12 tok/s
$2,299= 88 mo of VPSMac Studio (M5 Max, 36 GB)
36 GB · 460 GB/s · ~12–19 tok/s
$2,499= 96 mo of VPSGeForce RTX 5090 32 GB (card only)
Approximate32 GB · 1792 GB/s · ~47–73 tok/s
$3,440= 10+ yrs of VPSFramework Desktop (Ryzen AI Max+ 395, 128 GB)
Approximate128 GB · 256 GB/s · ~6.8–10 tok/s
$3,449= 10+ yrs of VPS
Buy · per month
$37.45
$36.08 hardware + $1.36 power
Rent · per month
$26.13
Leaseweb VPS 4 · 12 vCPU / 48 GB · Europe
Break-even
The Mac mini (M6, 32 GB) pays for itself after 52 months of replacing the VPS, and it streams an estimated 4–6.2 tok/s against the VPS's CPU-only pace.
How we count
On a CPU-only VPS Qwen2.5-Coder 32B needs about 23.4 GB of RAM: 19.9 GB of Q4_K_M weights, 2 GB of KV cache for an 8,192-token context and 1.5 GB for the OS and runtime. 93 of the 527 plans in the index fit and include a disk. A plan runs it comfortably with 20% headroom over the total, and is a tight fit with exactly enough.
Other quants: Q8_0 weights take 34.8 GB (near-lossless, about half of F16); F16 takes 65.5 GB.
The plans are ranked by estimated tokens per second per dollar, comfortable fits first. Speed is effective memory bandwidth divided by active weight bytes (4 GB/s per shared vCPU, 6 per dedicated), shown as a band. Real numbers depend on the host CPU generation, AVX-512 and AMX support and how noisy the neighbours are, so treat them as an order of magnitude.
Token prices are per million, blended 3 in : 1 out, before tax. A price below the maker's is listed only when the provider says why. A rented box costs the same whether or not you use it and runs whatever else you put on it; an API charges for what you send and nothing when you stop.
A GPU rate is per GPU for an hour, and the monthly figure is 730 hours of the whole instance at it, tax excluded. The token price is $0.10 per million, blended 3 in : 1 out, the cheapest price (Nscale), and the box is assumed busy every second of the month, which nothing is.
A machine tagged approximate is priced at the typical US street price in August 2026 rather than its maker's own price. Local speed is peak bandwidth times 0.7 efficiency (0.35 on CPU-only boards) divided by active weight bytes; unified-memory machines are assumed to give models 75% of their RAM. GPU cards need a host PC that is not in the price.
Frequently asked
- How much RAM does Qwen2.5-Coder 32B need?
- About 23.4 GB for CPU inference at Q4_K_M: 19.9 GB of weights, 2 GB of KV cache at 8,192 tokens of context, and 1.5 GB of headroom. Longer contexts need more KV cache (256 KB per token for this model).
- What is the cheapest VPS that can run Qwen2.5-Coder 32B?
- Leaseweb VPS 4 · 12 vCPU / 48 GB · Europe (48 GB RAM, 12 vCPU) at $26.13/mo excl. VAT runs it comfortably as of 31 Oct 2026. Leaseweb VPS 3 · 8 vCPU / 24 GB · Europe at $14.07/mo is a tight fit.
- How fast will Qwen2.5-Coder 32B run on a VPS without a GPU?
- Roughly 1.2–2.4 tok/s on the top pick. Token generation is limited by memory bandwidth, because each new token reads all of the weights once. More vCPUs help, dedicated ones most, but a GPU or an Apple silicon machine is 10 to 50 times faster.
Same lineup