Input cost clarity
Estimate the cost of instructions, user text, chat history, and retrieved context.
Token cost calculator
Token Cost Calculator
Calculate how input tokens, cached input tokens, and output tokens contribute to the cost of each AI API request.
Price requests from token components
Token cost is not one blended number. Many providers price input and output tokens separately, and some offer lower cached-input pricing for repeated prompt or context segments. This page focuses on the cost mechanics behind each request.
Calculator shortcut
Use the calculator with provider and model pricing to estimate input, cached input, output, and monthly cost.
Calculate token costsBenefits
Output cost planning
Model how generated answer length affects cost when output tokens carry a higher unit price.
Cached-input awareness
Account for cached input pricing where providers support discounted repeated context.
Providers and guides
Keep the calculator close to the source material: compare provider pages, then use short guides to refine token and usage assumptions.
All providers
Browse tracked AI providers and open provider-specific calculators.
OpenAI pricing
Review OpenAI models, source references, and calculator shortcuts.
Anthropic pricing
Compare Claude model pricing for similar token assumptions.
Gemini pricing
Open Gemini model pages and estimate provider-specific workloads.
Input vs output tokens
Understand how prompt and generated tokens affect API cost.
Reduce AI API costs
Use practical prompt, model, caching, and monitoring tactics.
Use cases
Per-request cost checks
Estimate one average request before multiplying it across traffic.
Prompt optimization
Measure how shorter instructions or smaller context windows can reduce input-token cost.
Response limits
Compare concise and verbose output settings before setting product defaults.
Pricing estimation warning
Provider pricing can change and may include special tiers, batch discounts, or terms not captured by a simple calculator.
Launch checklist
Make the estimate more useful
A few practical checks help developers and founders avoid surprises after real users arrive.
Common cost mistakes
Forgetting retries, long context, power users, and generated output length.
How to reduce AI API costs
Shorten prompts, cap output length, cache repeated answers, and route simple tasks to cheaper models.
Cheaper vs stronger models
Use stronger models when accuracy or reasoning changes the outcome; use cheaper models for routine work.
Before launching an AI feature
Ask who triggers requests, how often, how long responses are, and what happens during usage spikes.
Related calculators
AI Token Calculator
Estimate token usage before you ship AI features, then compare how the same prompt and response pattern changes cost across providers.
AI Usage Calculator
Estimate monthly AI API usage from traffic, request frequency, token volume, and model selection before growth makes costs harder to control.
LLM Cost Calculator
Compare LLM cost scenarios across providers and model tiers before choosing the API behind your product workflow.
FAQ
How are AI API costs calculated?
Most AI API estimates multiply input tokens, output tokens, and request volume by the selected model's token prices.
Why are output tokens more expensive?
Output tokens often cost more because the model is generating new text, which usually requires more inference work than reading input context.
What are cached input tokens?
Cached input tokens are repeated prompt or context tokens that some providers can reuse at a discounted price.