Price Per Token

Free

Compare LLM API pricing across providers. Includes token counter and cost calculator.

FreeFree tier

Type

Open Source

Visit Website GitHub (40 stars)

About Price Per Token

Price Per Token is a comprehensive, open-source platform for comparing LLM API pricing across over 300 AI models from providers like OpenAI, Anthropic, Google, Mistral, and more. It features an interactive grid with sortable columns for input/output/cached input costs, context windows, and benchmark scores (MMLU, GPQA). Users can filter models by capabilities such as tool use, vision, reasoning, web search, PDF input, caching, audio input, and open-source status. The site offers a weekly newsletter for pricing updates and a Model Context Protocol (MCP) server for integrating live pricing benchmarks directly into AI agents. Data is updated in real-time, with the latest refresh displayed on the homepage.

Key Features

Compare over 300 AI model costs from major providers (OpenAI, Anthropic, Google, Mistral, etc.)

Filter models by capability: Tool Use, Vision, Reasoning, Web Search, PDF Input, Caching, Audio In, Open Source

View detailed columns: context window, input/output/cached input pricing

Benchmark scores included (MMLU, GPQA) for performance comparison

MCP server for live LLM pricing benchmarks in AI agents

Weekly newsletter with pricing changes, new releases, and tools

Real-time updates with clear 'last updated' timestamp

Interactive grid with 'Try' and 'Vote' actions per model

Pros & Cons

Pros

Comprehensive comparison of 300+ models from many providers
Free and open-source with no paywall
Filters by specific capabilities streamline model selection
Includes benchmark scores for performance evaluation
Real-time pricing data with clear update frequency
MCP server enables direct integration into developer workflows

Cons

Primarily focuses on API pricing; does not include hosted or deployment costs
Benchmark scores may not reflect real-world performance for every use case
Sponsored listings (e.g., spot pricing) could influence visibility
Lacks direct integration with provider APIs for automatic cost calculation

Best For

Selecting cost-effective LLM APIs for application developmentComparing pricing across providers for budget optimizationResearching model capabilities and costs for AI projectsStaying informed about pricing changes via weekly newsletterIntegrating live pricing benchmarks into AI agents via MCP server

FAQ

How often is the pricing data updated?

The data is updated in real-time, with a 'last updated' timestamp displayed on the homepage (e.g., July 21, 2026 at 08:27 AM). A weekly newsletter also tracks pricing changes and new releases.

What providers are included?

Major providers such as OpenAI, Anthropic, Google, Mistral, and many others are included. The platform covers over 300 models as of the latest update.

Is the tool free to use?

Yes, Price Per Token is completely free and open-source. There is no paywall or subscription required.

Can I access the pricing data programmatically?

Yes, a Model Context Protocol (MCP) server is available for providing live LLM pricing benchmarks to AI agents and programs.

What filtering options are available?

You can filter models by capabilities such as Tool Use, Vision, Reasoning, Web Search, PDF Input, Caching, Audio In, and Open Source status.