~/runthismodel
daemon okbuild 5a3c91d00:00:00Z

Energy & cost

What does your local AI actually cost?

Pick your GPU and region. We’ll show monthly electricity, CO₂, and how it compares to just calling a hosted API.

Setup

Result

Energy / day
1.06 kWh
Energy / month
31.9 kWh
Electricity cost / month
$5.27
CO₂ / month
12.5 kg
Effective $/1 M output tokens
$0.290

Pure electricity cost per million tokens generated. Excludes hardware amortisation.

Vs. paying per token

Local cost: $0.290/1M tokens (pure electricity). Breakeven = tokens/month where local hardware pays off vs cloud.

DeepSeek V3

$0.27 in / $1.10 out per Mtok

$24.86/mo equiv.

breakeven at 3.8M tok/mo

Local saves $19.59/mo

Llama 3.1 8B (Groq)

$0.05 in / $0.08 out per Mtok

$2.36/mo equiv.

breakeven at 40.5M tok/mo

API cheaper by $2.91/mo

Gemini 2.0 Flash

$0.10 in / $0.40 out per Mtok

$9.07/mo equiv.

breakeven at 10.5M tok/mo

Local saves $3.80/mo

GPT-4o mini

$0.15 in / $0.60 out per Mtok

$13.61/mo equiv.

breakeven at 7.0M tok/mo

Local saves $8.34/mo

Claude Haiku 4.5

$0.80 in / $4.00 out per Mtok

$87.09/mo equiv.

breakeven at 1.1M tok/mo

Local saves $81.82/mo

GPT-4o

$2.50 in / $10.00 out per Mtok

$226.80/mo equiv.

breakeven at 422K tok/mo

Local saves $221.53/mo

"equiv." = what you’d pay the API to generate the same token volume your local hardware produces, assuming ~1:1 input:output tokens per chat. Breakeven = minimum tokens/month for local to be cheaper. Excludes hardware amortisation.

Don’t know what you can run? Check your hardware → · Full API price comparison →