The TokenCost MCP Server provides real-time LLM token pricing data and cost estimation tools for 60+ AI models across 15+ providers (OpenAI, Anthropic, Google, xAI, Meta, Mistral, DeepSeek, and more).

• Get Model Pricing (tokencost_get_model_pricing): Retrieve detailed pricing for a specific model (e.g., GPT-5, Claude Sonnet, Gemini Pro), including input/output costs per 1M tokens, context window size, and max output tokens.

• Compare Models (tokencost_compare_models): Side-by-side pricing comparison of 2–10 models simultaneously, showing cost differences to help choose the best option.

• Estimate Cost (tokencost_estimate_cost): Calculate the exact USD cost for a given number of input and output tokens with a specific model, broken down by input, output, and total.

• Find Cheapest Models (tokencost_find_cheapest): Discover the most cost-effective models, with optional filters by provider, minimum context window, and sorting by input, output, or combined cost.

• List All Models (tokencost_list_models): Browse all 60+ available models with IDs, names, and providers, optionally filtered by provider.

• List Providers (tokencost_list_providers): View all supported providers with their model counts and pricing ranges.

Pricing data is kept up-to-date within days of official changes and can be queried via natural language through AI assistants like Claude Desktop, Cursor, or Windsurf.

Provides real-time pricing data, cost estimation, and model comparisons for Amazon AI models like Nova.

Provides real-time pricing data, cost estimation, and model comparisons for Google AI models like Gemini.

Provides real-time pricing data, cost estimation, and model comparisons for Meta AI models like Llama.

Provides real-time pricing data, cost estimation, and model comparisons for NVIDIA AI models.

Provides real-time pricing data, cost estimation, and model comparisons for OpenAI AI models like GPT-4 and GPT-5.

Provides real-time pricing data, cost estimation, and model comparisons for Perplexity AI models.