AI Cost Calculator

Calculate Your AI Cost
Per User, Per Month

Know your AI unit economics before you scale. Compare 40+ models across OpenAI, Anthropic, Google, DeepSeek, and more.

OpenAI · $1.25/1M input · $10/1M output

~600 words

~300 words

Cost per request

$0.0050

AI cost per user / month

$0.250

1.3% of $20 revenue

Total monthly AI cost

$250.00

1,000 users × 50 req/mo

AI Gross Margin

98.8%

Revenue: $20.00kAI Cost: $250.00

Healthy margin for an AI-powered SaaS.

Track real costs with Orbit

Free tier · One-line SDK · No credit card

Cheaper Alternatives to GPT-5

Same token configuration, lower cost. Savings are per user per month.

Full Model Comparison

Cost per request with 800 input and 400 output tokens.

ModelProviderInput / 1MOutput / 1MCost / RequestCost / User / Mo
Llama 3.2 3BMeta$0.06$0.06$0.000072$0.0036
Gemini 2.0 Flash LiteGoogle$0.07$0.30$0.000180$0.0090
GPT-5 NanoOpenAI$0.05$0.40$0.000200$0.010
Mistral Small 3.2Mistral$0.10$0.30$0.000200$0.010
GPT-4.1 NanoOpenAI$0.10$0.40$0.000240$0.012
Gemini 2.5 Flash LiteGoogle$0.10$0.40$0.000240$0.012
Gemini 2.0 FlashGoogle$0.10$0.40$0.000240$0.012
GPT-4o MiniOpenAI$0.15$0.60$0.000360$0.018
Grok 4.1 FastxAI$0.20$0.50$0.000360$0.018
DeepSeek ChatDeepSeek$0.28$0.42$0.000392$0.020
Grok 2 MinixAI$0.20$1.00$0.000560$0.028
DeepSeek V3DeepSeek$0.27$1.10$0.000656$0.033
Claude 3 HaikuAnthropic$0.25$1.25$0.000700$0.035
GPT-4.1 MiniOpenAI$0.40$1.60$0.000960$0.048
GPT-5 MiniOpenAI$0.25$2.00$0.0010$0.050
Mistral Large 3Mistral$0.50$1.50$0.0010$0.050
Llama 3.3 70BMeta$0.88$0.88$0.0011$0.053
Gemini 2.5 FlashGoogle$0.30$2.50$0.0012$0.062
Gemini 3 FlashGoogle$0.50$3.00$0.0016$0.080
o3 MiniOpenAI$1.10$4.40$0.0026$0.132
o4 MiniOpenAI$1.10$4.40$0.0026$0.132
Claude Haiku 4.5Anthropic$1.00$5.00$0.0028$0.140
Claude 3.5 HaikuAnthropic$1.00$5.00$0.0028$0.140
GPT-4.1OpenAI$2.00$8.00$0.0048$0.240
o3OpenAI$2.00$8.00$0.0048$0.240
GPT-5.1OpenAI$1.25$10.00$0.0050$0.250
GPT-5OpenAI$1.25$10.00$0.0050$0.250
Gemini 2.5 ProGoogle$1.25$10.00$0.0050$0.250
Grok 2xAI$2.00$10.00$0.0056$0.280
GPT-4oOpenAI$2.50$10.00$0.0060$0.300
Gemini 3 ProGoogle$2.00$12.00$0.0064$0.320
GPT-5.2OpenAI$1.75$14.00$0.0070$0.350
Claude Sonnet 4.5Anthropic$3.00$15.00$0.0084$0.420
Claude Sonnet 4Anthropic$3.00$15.00$0.0084$0.420
Grok 4xAI$3.00$15.00$0.0084$0.420
Claude Opus 4.5Anthropic$5.00$25.00$0.014$0.700
Claude Opus 4.1Anthropic$15.00$75.00$0.042$2.10
o3 ProOpenAI$20.00$80.00$0.048$2.40
GPT-5.2 ProOpenAI$21.00$168.00$0.084$4.20

Prices per 1M tokens from official provider pricing pages. Click a row to select that model.

Questions

How accurate are these calculations?

Prices are from official provider pricing pages as of January 2026. Actual costs may vary with volume discounts, batch API pricing, or cached tokens. This gives you a reliable baseline estimate.

What are input vs output tokens?

Input tokens are what you send to the model (prompts, context, system messages). Output tokens are what the model generates back. As a rough guide, 1 token is about 0.75 words in English.

Why does cost per user matter?

If you charge $20/month per user but your AI costs $8/month per user, your AI gross margin is only 60%. Understanding cost per user helps you price correctly and stay profitable as you scale.

How do I track real costs in production?

This calculator gives estimates. For real-time, per-feature cost tracking from actual API calls, integrate Orbit's SDK. It takes one line of code and tracks every request automatically.

Estimates are good.
Real data is better.

Orbit tracks actual AI costs from your production API calls. One-line SDK, real-time dashboards, per-feature breakdown.

Start tracking free