What it does
Price Per Token is a comparison table of what language model APIs cost. It covers more than 300 models from OpenAI, Anthropic, Google, Mistral and other providers, and lists context length, input price, output price and cached-input price for each one. The table can be filtered by capability, including tool use, vision, reasoning, web search, PDF input, caching, audio input and open source. Benchmark columns for coding, MMLU and GPQA appear where scores exist, and a separate leaderboard page covers benchmarks. Each row has Try and Vote actions, and a page tracks new model releases. The site also announces an MCP that lets AI agents read live pricing and benchmark data. A weekly newsletter on pricing changes, new releases and tools is offered, and the home page carries a sponsored placement for a spot-priced API. The site does not state whether it has paid plans or any account requirement.
Capabilities
- Input, output and cached-input prices for 300+ models in one table
- Filters for capabilities such as vision, reasoning and tool use
- Benchmark columns and a benchmark leaderboard page
- A page of new model releases
- An MCP that gives AI agents live pricing and benchmarks
When to use it
- You are choosing a language model API and want to compare costs
- You need a model with a specific capability, such as vision or PDF input, at a low price
- You want an agent to look up current model prices through MCP
- You want to keep an eye on new model releases and price changes
