How to Add Cost-Aware Model Selection to Your AI Agent

WhichModel is an open MCP server for cost-aware model selection.

It tracks pricing and capabilities across 100+ LLM models.

Add it to an MCP client with https://whichmodel.dev/mcp; no API key or installation is needed.

It can return a recommended model, a budget alternative, cost estimates, and reasoning.

Pricing updates every 4 hours, and at 10,000 calls/day the gap is $216/day between $15/M-token and $0.60/M-token models.