# llmcalc.app

> Verified LLM API pricing and cost decision tools. 16 current models across Anthropic, OpenAI, Google, and Groq, each verified against the provider's official pricing page with a dated stamp (last refresh 2026-06-12). Includes a Claude Code plan verdict tool that computes whether Pro, Max, or pay-as-you-go API is cheaper from a user's real usage.

Key facts:
- All prices are USD per 1M tokens and carry a last_verified date.
- Machine-readable data: https://llmcalc.app/api/pricing.json (free, attribution requested).
- The Claude Code verdict tool parses ccusage output entirely in the browser; usage data is never uploaded.
- Models removed from a provider's public price list are flagged "delisted" and kept for reference only.

## Tools

- [Claude Code plan verdict](https://llmcalc.app/claude-code-cost): paste ccusage output, get the API-equivalent bill, value multiple, plan verdict, and a shareable receipt image
- [Agent cost simulator](https://llmcalc.app/agent-calculator): multi-turn agent loop costs with tool calls, retries, and context growth
- [Prompt caching ROI](https://llmcalc.app/cache-calculator): cache write vs read vs hit-rate break-even
- [Batch decision tool](https://llmcalc.app/batch-calculator): when the 50% batch discount fits a workload
- [Generic cost calculator](https://llmcalc.app/calculator): per-1M-token comparison across all tracked models
- [Token counter](https://llmcalc.app/counter): multi-model token counting

## Pricing data

- [All tracked models](https://llmcalc.app/models/): table with verification status
- [Machine-readable JSON](https://llmcalc.app/api/pricing.json): full pricing dataset including cache tiers, batch discounts, and long-context tiers
- [Changelog](https://llmcalc.app/changelog): dated log of every price change and verification pass

## Optional

- [About](https://llmcalc.app/about): what this site is and is not
- [Full pricing snapshot](https://llmcalc.app/llms-full.txt): current per-model prices inline