Hey everyone 👋 Sylvain here. Every LLM cost calculator I tried wanted something from me first — an email, an account, my usage data. I just wanted the division. So I built the arithmetic and nothing else: you enter token counts and per-million rates, it returns the bill, in your browser. No backend involved — no API calls, no analytics, no cookies. Your numbers never leave your tab. The more I modeled real workloads, the more I realized the formula was never the problem: Code block input tokens × input rate + output tokens × output rate, all / 1,000,000 Every surprise in an LLM invoice is a rule outside that formula — cached input discounts, tier boundaries, retries, minimum purchases. That's why every page on the site lists what the estimate excludes instead of quietly hiding it. That exclusions list is honestly the feature I'm proudest of. What's inside today: • 65 model pages (live rates from a public catalogue) • 60+ workload calculators — coding-agent sessions, RAG pipelines, chat bots, monthly budgets • Model-vs-model comparison pages that show the method (rates rot; math doesn't) • MIT-licensed, source public Two things I'd love from you: What's missing from the exclusions list? (It grows from feedback — cache rules and retry storms came from users.) What workload shape should the next calculator model? Agents on long contexts? Batch pipelines? Thanks for checking it out — happy to answer anything about the method, the math, or why I refuse to print prices in static pages 😄 https://rynler.com/calculators

Rynler Token Cost Calculator

LLM cost calculator. No email. No tracking. Just math.

LLM cost calculator. No email. No tracking. Just math.