@calcis/pricing

Open dataset of LLM API pricing for 25+ models across OpenAI, Anthropic, and Google.

Live pricing for 25+ models, side-by-side comparisons, and a web estimator: https://calcis.dev

• Full price index: https://calcis.dev/models
• Compare models: https://calcis.dev/compare
• API reference: https://calcis.dev/api-docs

Shipped as a single JSON file so you can:

• Vendor it into a build,
• Import it at runtime,
• Use it in a script that doesn't even have Node installed.

The source of truth is the Calcis application's pricing table. Every row is re-verified against the provider's rate card on a rolling schedule; the lastUpdated field is the ISO date a human last confirmed it.

Install

npm install @calcis/pricing

Use

// Node / bundler
import pricing from "@calcis/pricing";

for (const m of pricing.models) {
  console.log(m.displayName, m.inputPricePerMTok, m.outputPricePerMTok);
}

// Fetch directly in the browser
const res = await fetch(
  "https://unpkg.com/@calcis/pricing/index.json"
);
const pricing = await res.json();

No build step, no dependencies. index.json is the entry point.

Schema

interface PricingFile {
  version: string;           // semver of the dataset shape
  generatedAt: string;       // ISO date the file was generated
  source: string;            // canonical URL for richer data
  note: string;              // unit + field conventions (human-readable)
  models: ModelPricing[];
}

interface ModelPricing {
  /** Exact API model id to pass to the provider SDK. */
  id: string;
  /** Human-friendly name (e.g. "Claude Opus 4.7"). */
  displayName: string;
  /** "openai" | "anthropic" | "google". */
  provider: "openai" | "anthropic" | "google";
  /** USD per 1,000,000 input tokens. */
  inputPricePerMTok: number;
  /** USD per 1,000,000 output tokens. */
  outputPricePerMTok: number;
  /** Context window in tokens. */
  contextWindow: number;
  /** Max output tokens per response (omitted when provider doesn't publish). */
  maxOutput?: number;
  /** ISO date a human last verified the row against the provider. */
  lastUpdated: string;
}

What's NOT in here (by design)

This package is deliberately the simple case: base input / output prices and context window. It does not ship:

• Cached-input pricing (OpenAI + Gemini)
• Anthropic cache-write pricing
• Gemini long-context surcharge tiers
• Tokenizer identity or multipliers
• Batch / async discounts

Those live in the full Calcis app where they can be modelled against the cost estimator. If you have a use case that needs them as a dataset too, open an issue.

License

MIT. See LICENSE.

Sources

Canonical rate cards the data is derived from:

• Anthropic: https://platform.claude.com/docs/en/about-claude/models/overview
• Google Gemini: https://ai.google.dev/gemini-api/docs/pricing
• OpenAI: https://developers.openai.com/api/docs/models

Links

• Website: https://www.calcis.dev
• Upstream repo: https://github.com/rc397/Calcis
• Full estimator package: @calcis/cost-estimator