AI Model Cost Calculator

What will AI actually cost your business?

Estimate monthly inference spend across today’s leading models. Adjust the inputs — the numbers update as you go.

Your workload

Requests / month50K/mo
10010K1M10M

Average tokens per request. Defaults are typical for the selected use case.

Total tokens / month38M

Estimated monthly cost

Sorted cheapest first · USD
  • GPT-5.6 LunabudgetCHEAPEST
    $20.00
    $0.20 in · $1.20 out / 1M$0.00040 / request
  • DeepSeek V4 Probudget
    $21.75
    $0.43 in · $0.87 out / 1M$0.00044 / request
  • Claude Haiku 4.5budget
    $87.50
    $1.00 in · $5.00 out / 1M$0.00175 / request
  • Gemini 3.6 Flashmid
    $131
    $1.50 in · $7.50 out / 1M$0.00263 / request
  • Claude Sonnet 5premium
    $175
    $2.00 in · $10.00 out / 1M$0.00350 / request
  • Gemini 3.1 Propremium
    $200
    $2.00 in · $12.00 out / 1M$0.00400 / request
  • GPT-5.6 Solpremium
    $500
    $5.00 in · $30.00 out / 1M$0.010 / request
  • TYPICAL ROUTED SETUP80% GPT-5.6 Luna · 20%
    $51.00
    Cheapest budget tier + your chosen premium model, blended$0.00102 / request
P
Our take

The flagship models here cost roughly 25.0× the efficient tier for this workload. Most teams start on an efficient model, measure quality on real traffic, and escalate only the requests that need it — a routing pattern that usually saves 40–70%.

Want a real estimate for your use case?

These are back-of-envelope numbers. We build production AI systems — routing, caching, evals, and the infrastructure around them — and can model your actual workload.