Skip to content

README.md

Latest commit

 

History

History
138 lines (94 loc) · 3.86 KB

README.md

File metadata and controls

138 lines (94 loc) · 3.86 KB
Raw

tokenpricing (TypeScript SDK)

npm version

API pricing math for 1k+ AI models from the canonical tokenpricing database with multi-currency and cache-token pricing support.

Why tokenpricing?

Token pricing for AI models changes frequently across different providers and model types. This library now consumes the canonical database published from this repository, synchronized directly from upstream pricing sources every six hours.

Important: This library does not estimate token counts from strings or messages. tokenpricing focuses solely on providing accurate, current pricing data.

Features

  • Up-to-date AI model pricing from the tokenpricing canonical database
  • Caching with 6-hour TTL for pricing data
  • Multi-currency conversion via JSDelivr currency API with a 24-hour cached USD rates map
  • TypeScript-first with full type definitions
  • ESM and CJS dual-format package
  • Zero dependencies beyond fuse.js (for fuzzy matching)

Installation

pnpm add @atenareply/tokenpricing

Or with npm:

npm install @atenareply/tokenpricing

Usage

import { getPricing, computeCost } from "@atenareply/tokenpricing";

Get Pricing

const pricing = await getPricing("openai/gpt-5.2");
console.log(`Input: $${pricing.inputPerMillion.toFixed(2)}/1M tokens`);
console.log(`Output: $${pricing.outputPerMillion.toFixed(2)}/1M tokens`);

Get Pricing in Another Currency

const pricing = await getPricing("openai/gpt-5.2", "EUR");
console.log(`Input: €${pricing.inputPerMillion.toFixed(2)}/1M tokens`);

Compute Cost

const cost = await computeCost("openai/gpt-5.2", 1000, 500, "EUR", { cacheReadTokens: 250, cacheCreationTokens: 100 });
console.log(`Total cost: €${cost.toFixed(6)}`);

Helpful Error Messages

When you make a typo in a model ID or currency code, tokenpricing provides helpful "Did you mean?" suggestions:

await getPricing("openai/gpt4");
// Error: Model not found: openai/gpt4. Did you mean 'openai/gpt-4'?

await getPricing("openai/gpt-4", "ERU");
// Error: Unsupported currency: ERU. Did you mean 'EUR'?

API

getPricing(modelId, currency?)

Get pricing info for a specific model.

  • modelId — Model identifier (e.g., "openai/gpt-4")
  • currency — Target currency code (default: "USD")
  • Returns Promise<PricingInfo>

computeCost(modelId, inputTokens, outputTokens, currency?, options?)

Compute total cost for a specific model given token counts.

  • modelId — Model identifier
  • inputTokens — Number of input tokens
  • outputTokens — Number of output tokens
  • currency — Target currency code (default: "USD")
  • options.cacheReadTokens / options.cacheCreationTokens — optional cached-token usage
  • Returns Promise<number>

PricingInfo

interface PricingInfo {
  inputPerMillion: number;
  outputPerMillion: number;
  cacheReadPerMillion?: number;
  cacheCreationPerMillion?: number;
  currency: string;
}

Data Source

Pricing data is sourced from the canonical tokenpricing dataset generated in this repository from OpenRouter and LiteLLM.

Caching uses a 6-hour TTL aligned to the canonical sync cadence. Currency conversion uses daily USD base rates from the JSDelivr currency API with a 24-hour cache.

Development

Setup

pnpm install

Commands

pnpm build          # Build ESM + CJS bundles
pnpm test           # Run tests
pnpm test:coverage  # Run tests with coverage
pnpm lint           # Lint and format check
pnpm lint:fix       # Auto-fix lint/format issues
pnpm typecheck      # TypeScript type check

Credits

License

See LICENSE file for details.