Quick start
-
Select model
Includes GPT, Claude, Gemini reference rates.
-
Enter tokens
Set estimated input and output token counts.
LLM Pricing Calculator
Pick a model and token counts to estimate USD cost per API call. Useful for budgeting, model comparison, and cost-aware design.
Privacy: processed locally, never uploaded.
↓ Paste in the input area below to see results instantly
Prices are indicative; check each provider for official rates.
Est. cost (USD)
$0.0027
Input price
$0.15/1M
Output price
$0.6/1M
Notes
Billing note
Estimates cost per request from per-million token rates. Actual bills may include caching, batch pricing, etc.
Pick a model and token counts to estimate USD cost per API call. Useful for budgeting, model comparison, and cost-aware design.
Price source
Rates are indicative; bills may include caching, batch pricing, or regional differences.
Features and use cases
Estimate per-call and monthly LLM cost from token counts and model pricing with side-by-side compare.
Use for product pricing, model selection, budget approvals, and cost anomaly investigation.
Typical Workflow
1. Select Model: Choose target LLM (e.g. GPT-4 or Claude 2) - prices vary up to 10x between models. 2. Estimate Tokens: Test sample inputs in playground or use rule of thumb (English: 1 Token≈1.3 chars, Chinese≈2 chars). 3. Calculate single-call cost, then multiply by daily calls for monthly estimates.
Pro Tip: Compare multiple models side-by-side. Example: Evaluate GPT-3.5 vs Llama 2 for long-text scenarios, or design tiered strategies (simple queries→cheap model, complex tasks→premium model). Adjust tokens live to see price sensitivity.
Examples
Example
Input
gpt-4o-mini, 10k in / 2k out
Output
~$0.0027 USD
FAQ
Image tokens?
Text tokens only; check provider docs for multimodal.
Are system prompts included in token count?
Yes, system prompts (e.g. role-setting instructions) count as input tokens. However, some platforms like OpenAI have special billing rules for system prompts - recommend testing with actual API.