LLM cost calculator: cloud API vs local GPU

For a given workload, estimate monthly cost on a cloud API offering and the cost of running an equivalent open-weight model on your own GPU hardware, with amortisation, energy and labour made explicit. Not a benchmark: local throughput is estimated from hardware specifications with efficiency factors you can override below. All amounts in USD. EUR display

Fairness controls

Quality impact of fp8 quantisation is not modelled; this comparison is cost only.

Amortisation horizon
Workload
Or compute from requests/day
Cloud

$21.46 / month ($2.15 / Mtok effective)

$0.60 / 1M chars (o200k, blended slice)

Breakdown
Input$5.54
Output$15.84
Cache read$0.0792
Cache write$0.00
Batch saving-$0.00
Per-request fees$0.00
Other hosts of this model
Host$/month$/Mtok
DeepSeek V4 Pro (NVIDIA)$5.31$0.53
DeepSeek V4 Pro (OpenRouter)$10.23$1.02
DeepSeek V4 Pro (Relace, fp4)$10.23$1.02
DeepSeek V4 Pro (SiliconFlow, fp8)$10.44$1.04
DeepSeek V4 Pro (DeepSeek, off-peak)$10.73$1.07
DeepSeek V4 Pro (Fireworks AI)$10.92$1.09
DeepSeek V4 Pro (DeepInfra, flex+fp8)$12.83$1.28
DeepSeek V4 Pro (DigitalOcean)$13.11$1.31
DeepSeek V4 Pro (Cloudflare Workers AI)$15.39$1.54
DeepSeek V4 Pro (Parasail, fp8)$15.99$1.60
DeepSeek V4 Pro (DeepInfra, fp8)$16.04$1.60
DeepSeek V4 Pro (DeepInfra)$16.04$1.60
DeepSeek V4 Pro (Alibaba Cloud, fp8)$17.49$1.75
DeepSeek V4 Pro (Reka AI)$18.84$1.88
DeepSeek V4 Pro (Alibaba Cloud, fp8)$19.27$1.93
DeepSeek V4 Pro (Novita AI, fp8)$19.76$1.98
DeepSeek V4 Pro (Novita AI)$19.76$1.98
DeepSeek V4 Pro (Venice)$20.73$2.07
DeepSeek V4 Pro (Atlas Cloud, fp4)$20.81$2.08
DeepSeek V4 Pro (Baidu, fp8)$20.87$2.09
DeepSeek V4 Pro (StreamLake, fp8)$21.49$2.15
DeepSeek V4 Pro (GMI Cloud, fp8)$21.49$2.15
DeepSeek V4 Pro (NextBit, fp8)$21.49$2.15
DeepSeek V4 Pro (Azure AI Foundry)$21.49$2.15
DeepSeek V4 Pro (Crusoe, fp8)$21.50$2.15
DeepSeek V4 Pro (Nebius AI Studio)$21.60$2.16
DeepSeek V4 Pro (Nebius AI Studio, fp8)$21.60$2.16
DeepSeek V4 Pro (Azure AI Foundry)$23.63$2.36
DeepSeek V4 Pro (DeepInfra, fp8+priority)$24.06$2.41
Local hardware
Custom hardware (used only when Hardware = Custom)
Cost assumptions
Advanced: measured throughput overrides
Cloud$21.46 / mo
Local, workload$– / Mtok
Local, capacity$– / Mtok
Cheapercloud

Breakeven and sensitivity

Breakeven volume: – tokens/month. Breakeven utilisation: –.

Monthly cost vs. tokens/month

Cloud cost scales linearly with volume; local cost steps up each time another box is needed. Log scale on tokens/month.

Assumptions

KeyValueSourceOverridden
hoursPerMonth730Units: a month is 730 hours (8,760 hours per year divided by 12)no
daysPerMonth30.4Workload: 30.4 days per month when converting requests per dayno
weightsOverheadFactor1.05Memory model: weights take 5% more memory than their raw sizeno
fixedOverheadGb2Memory model: 2 GB reserved for runtime and buffersno
tpEffPcie0.85Memory model: tensor-parallel efficiency over PCIeno
tpEffNvlink0.95Memory model: tensor-parallel efficiency over NVLinkno
tpEffUnified1Memory model: no tensor-parallel loss on a single unified-memory machineno
defaultContextTokens81928192 tokens; typical chat context, editableno
decodeEff0.7Throughput model: decode reaches 70% of memory-bandwidth speedyes
prefillEff0.5Throughput model: prefill reaches 50% of peak computeyes
batchEffK0.04Throughput model: batching efficiency 1 / (1 + 0.04 (n - 1))yes
utilisation0.3Throughput model: hardware busy 30% of the month, editableno
maxUtilisation0.95Throughput model: a box is never loaded above 95%no
sparesPct0.1Monthly cost: 10% added to hardware cost for sparesno
residualPct0.2Monthly cost: 20% of the GPU price recovered at the end of the horizonno
loadFactor0.8Monthly cost: GPUs draw 80% of rated power under loadyes
horizonMonths36Monthly cost: hardware written off over 36 months (24 and 48 offered)no
labourHours2Monthly cost: 2 hours of upkeep per month, editableno
labourRate60Monthly cost: $60 per hour of upkeep, editableno
hostingMonth0Monthly cost: no hosting or colocation fee unless you enter oneno
softwareMonth0Monthly cost: no software or support fee unless you enter oneno
electricityUsdKwh0.1453US commercial average retail price, EIA Electric Power Monthly table 5.6.A, July 2026 (14.53 cents per kWh)no
latencyTargetTps20Workload: an interactive stream should reach 20 tokens per secondno
tokenizerSliceblendedTokenizer normalisation: blended average over the measured corpus slicesyes