Guides
Price guides and calculators
LLM API prices from each provider's own page, the same workload priced on each model, free cost calculators, and plain comparisons of AI routers and gateways. No sign up to read any of it.
Price guides
List prices per million tokens, dated and sourced, and what a real workload costs on each model.
9 pages
- LLM API pricing in 2026: every major model, per tokenInput, cached input, and output prices for Claude, GPT, Gemini, and xAI models, with sources and dates.
- How to cut your OpenAI API costsPick the smallest GPT model that does the job, keep cached input working, batch what can wait, and cap each key.
- How to cut your Anthropic (Claude) API costsOpus, Sonnet, or Haiku, cache markers that pay off, the Batch API, and where long prompts cost more.
- Cheapest LLM API for coding, priced per taskWhat a month of coding requests costs on each model, why cached input matters most here, and when to pay for a bigger model.
- Cheapest LLM API for summarization, worked out10,000 documents priced on every model, with and without batch, and how to check a cheap model is good enough.
- Cheapest LLM API for classification and extractionA million labels and 50,000 extracted records priced per model. Small models usually win here.
- GPT vs Claude API pricing, model by modelOpenAI and Anthropic models side by side at each size, the same workload priced on both, and the fine print.
- Claude vs Gemini API pricing, model by modelAnthropic and Google models side by side, launch prices and when they end, and long prompt prices.
- GPT vs Gemini API pricing, model by modelOpenAI and Google models side by side, launch prices and when they end, and the same workload priced on both.
Cost calculators
Free, in your browser, no sign up.
2 pages
Routers, gateways, and cutting costs
How routing works, how the options compare, and what lowers a bill.
6 pages
- How to reduce your LLM API costsRouting, prompt caching, caps per key, batch, and sending less. Plus when routing does not help.
- AI API router: how it works and how the options compareWhat OpenRouter, Requesty, Portkey, LiteLLM, and FUSR.ai add to your bill, and how fusr/auto picks a model.
- Use your own API key in Cursor, and what it costsWhat your own key covers in Cursor, what Cursor still charges, and the setup steps.
- Hosted LiteLLM: the proxy, without running itFUSR.ai runs on LiteLLM. What it adds, and when running LiteLLM yourself is the better call.
- OpenRouter alternative with your own keys and no markupOpenRouter's fees and BYOK terms next to FUSR.ai and self hosted LiteLLM, and when each one fits.
- LiteLLM vs Portkey vs OpenRouter: which AI gateway fitsWho runs the proxy, what each adds to your bill, and what you get for caching, caps, and logs.
Setup
Get FUSR.ai working in your tools.
1 page