🧮 LLM Cost Calculator

A powerful Node.js library to calculate token counts and estimated costs for Large Language Models (LLMs).

✨ Features

• 🎯 Accurate token counting for various LLM models
• 💰 Real-time cost estimation using up-to-date pricing from OpenRouter API
• 🔄 Auto-fetches latest model prices from OpenRouter API
• 🔌 Works offline with built-in pricing data
• ⚡ Fast and efficient with caching support

📦 Installation

npm install llm-cost-calculator

🚀 Quick Start

import { getEstimatedCost } from 'llm-cost-calculator';

// Calculate cost for a conversation with realtime prices
const result = await getEstimatedCost({
  model: 'anthropic/claude-3.5-sonnet',
  input: 'Write a function to calculate fibonacci numbers.',
  output: `Here's a recursive function to calculate Fibonacci numbers:

function fibonacci(n: number): number {
  if (n <= 1) return n;
  return fibonacci(n - 1) + fibonacci(n - 2);
}`
//   options: {
//     offline: true // Pass this option to use local pricing data
//   }
});

console.log(result);
// Output:
// {
//   inputTokens: 9,
//   outputTokens: 42,
//   cost: 0.000153 // Cost in USD
// }

🛠️ API Reference

getEstimatedCost(params)

Main function to calculate tokens and cost.

Parameters:

• model (string): Name of the LLM model
• input (string, optional): Input/prompt text
• output (string, optional): Output/completion text
• options (object, optional):
  • offline (boolean): Run in offline mode using local pricing data
  • timeout (number): API request timeout in milliseconds

Returns:

{
  inputTokens: number;  // Number of tokens in input
  outputTokens: number; // Number of tokens in output
  cost: number;        // Estimated cost in USD
}

🌐 Supported Models

Supports a wide range of models including:

OpenAI Models

• openai/gpt-4 - GPT-4 base model
• openai/gpt-3.5-turbo - Fast and cost-effective
• openai/gpt-3.5-turbo-0125 - Latest GPT-3.5 with 16k context

Anthropic Models

• anthropic/claude-3.5-sonnet - Latest Claude 3.5 Sonnet
• anthropic/claude-3-sonnet - Claude 3 Sonnet base
• anthropic/claude-3.5-sonnet:beta - Self-moderated version

Google Models

• google/gemini-2.0-flash-thinking-exp:free - Experimental Gemini with thinking process
• google/palm-2-chat-bison - PaLM 2 for general chat
• google/palm-2-codechat-bison - PaLM 2 specialized for code

Meta/Llama Models

• meta-llama/llama-2-13b-chat - Llama 2 13B chat model
• mistralai/codestral-2501 - Specialized for coding tasks
• microsoft/phi-4 - Efficient 14B parameter model

Other Notable Models

• deepseek/deepseek-chat - DeepSeek V3 for general chat
• qwen/qvq-72b-preview - Visual reasoning specialist
• minimax/minimax-01 - Text and image understanding

All models are available here

Pricing data is automatically fetched from OpenRouter API, with fallback to local pricing data. The library caches pricing data to minimize API calls while ensuring you always have access to the latest rates.

Note: Model availability and pricing may vary. Check the OpenRouter API for the most up-to-date list and pricing information.

📝 License

MIT

🤝 Contributing

Contributions are welcome! Feel free to:

1. Open issues for bugs or feature requests
2. Submit pull requests
3. Improve documentation
4. Share your use cases

💡 Tips

• Use offline mode when you don't need real-time pricing updates
• Cache results for repeated calculations
• Check token counts before making API calls to estimate costs