LLM Prices is a developer-oriented, machine-readable LLM API pricing data project maintained by modelpricewatch.com, with its repository on GitHub (romanshumy/llm-prices-data). Its goal is to provide a unified data source for tools, applications, and research that need accurate, timely, and traceable model invocation cost data.
Purpose The project primarily addresses the problems of fragmented, rapidly changing, and inconsistently defined LLM API pricing: it provides structured data for scenarios that need to compare costs across models, build cost estimation tools, monitor price changes, or cite authoritative pricing figures.
Main capabilities
- Covers 40 providers and 222 active models, including OpenAI, Anthropic, Google, Mistral, xAI, DeepSeek, Meta, Groq, Together, and others.
- Each price links to the provider's official pricing page for traceability and verification.
- Cross-checked daily against independent reference sources; top models are re-read from provider pages weekly with dated evidence.
- The repository updates automatically after each refresh cycle; the last sync time is 2026-10-14.
- Provides a live API that requires no API key and supports CORS, making it convenient for direct calls from frontends or servers.
- Provides an Atom feed for price changes, making it easy to track price adjustments.
- Publishes The LLM Price Index: a fixed, equal-weight basket of frontier flagship models, blended at a 3:1 input-output ratio, re-verified daily, used to give a single citable figure for the "cost of frontier intelligence per million tokens."
How to use Developers can obtain data files through the GitHub repository or directly call its public live API (no key required, CORS open). The API documentation is at https://modelpricewatch.com/api/ , the price change feed is https://modelpricewatch.com/news/feed.xml . The methodology description for The LLM Price Index is at https://modelpricewatch.com/price-index/methodology/ . The repository also provides MCP bridge code under the MIT license.
Use cases
- Build model cost comparison, routing, or budgeting tools.
- Cite traceable, verifiable pricing figures in applications or documentation.
- Monitor LLM API price changes and trigger notifications.
- Cite The LLM Price Index in research or reporting as a unified cost metric.
Limitations and notes
- Historical data points are labeled by source: self-verified /snapshot captures and litellm-archive (rebuilt from the git history of the open-source LiteLLM repository) are clearly distinguished and are not silently mixed.
- Deeper 2024→2026 historical backfill is still in progress, and some models currently have short timelines.
- Known baseline discontinuity issues are being audited.
- Data is released under the CC-BY-4.0 license and may be used free of charge for any purpose, but attribution to modelpricewatch.com is required; the MCP bridge code uses the MIT license.
