Free LLM Pricing API — AI Model Costs, Specs & Benchmarks

CloudPrice

2 min read Original article ↗

CloudPrice AI Models API

One API for LLM specs, prices and cost calculation

Pick any model, choose an operation and copy the exact REST URL for your integration.

Available as

Free TierPublic endpoints with no auth required.

Pro TierHigher request limitsGuaranteed throughputBatch exportsDedicated supportContact us

Explore specs, prices and usage cost

GEThttps://ai.cloudprice.net/api/v1/models/anthropic-claude-5-1-fable/pricing/calculate?input_tokens=1500&output_tokens=500

Send token and cache usage and get the cheapest matching provider plus every option ranked by total cost, with a per-dimension breakdown for each.

Recipes

Six common jobs you can solve with one URL — paste into curl, your app or your browser.

GET/models/{id}/pricing/calculate

Find the cheapest provider for a model

You have a token budget and want the lowest total cost across every provider that hosts the same model — including alternative tiers like batch and flex.

GET/litellm_model_prices.json

🚅 Use as your LiteLLM cost map

Drop-in replacement for litellm's model_prices_and_context_window.json, refreshed daily from each provider's own rates. Point LITELLM_MODEL_COST_MAP_URL at it and get models and providers litellm doesn't ship yet.

GET/calculate

Calculate cost across many models at once

Same workload, up to 5 candidate models, one request. Get a ranked list of total costs without writing a loop.

GET/models/{id}/compare

Compare flagship models head-to-head

Pick between flagship models. Get pricing for each across the same dimensions in one request.

GET/models/{id}/similar

Find a cheaper alternative with similar quality

Your current model works, but a comparable cheaper option may exist from another vendor — same capabilities, lower bill.

GET/benchmarks/{slug}/leaderboard

Pull a benchmark leaderboard

Need a ranked list of top models on a specific benchmark — handy for comparison pages, blog posts or CMS cards.

Need the full reference? All 17 endpoints with every parameter and response shape live in the interactive OpenAPI spec or as raw openapi.json.