LLM API Prices

Ask your AI assistant what an LLM API costs

This site runs a free MCP server: add it to Claude Code, Cursor, VS Code or Claude, then ask where a model is cheapest or what your workload would cost, and the assistant answers from the same prices as this site: 521 models, every seller we track, refreshed every 6 hours. No key and no sign-up.

https://easiapiai.com/mcp

Add it

Claude Code

claude mcp add --transport http easiapiai https://easiapiai.com/mcp

Cursor

In ~/.cursor/mcp.json (or .cursor/mcp.json in a project):

{
  "mcpServers": {
    "easiapiai": { "url": "https://easiapiai.com/mcp" }
  }
}

VS Code

In .vscode/mcp.json:

{
  "servers": {
    "easiapiai": { "type": "http", "url": "https://easiapiai.com/mcp" }
  }
}

Claude (web and desktop)

Settings → Connectors → Add custom connector, and paste https://easiapiai.com/mcp.

Anything else

Any client that speaks MCP over Streamable HTTP: point it at https://easiapiai.com/mcp.

Ask it

Tools

search_models
Find models by name, developer, minimum Intelligence Index score or maximum input price, sorted by relevance, score or price. Each result carries its cheapest standard price and seller.
get_model_prices
Every seller of one model, cheapest standard price first, with the official list price and the batch, flex, priority, off-peak and promo prices labelled as such.
estimate_cost
The monthly bill for your workload (tokens per request, requests per day, share of input served from cache) on up to ten models, each at its cheapest seller, and with one model, on every seller.

Good to know

Read-only, and it only ever returns what this site publishes: the cheapest standard price is the headline, and promos are labelled because a seller can end them any day. Every answer links the model's page here. Prices change: check the seller's own page before you commit spend. Last refresh 2026-10-15 16:51:05 UTC.