What Online Model Cost is
Online Model Cost is a browser set of tools for AI model cost. Use it as an LLM cost calculator when you need a number before a prompt ships, and again when you need a sketch of a monthly bill. The site does not replace a vendor invoice. It prices catalog rates for Claude, GPT, Gemini, and Grok so you can run an LLM price check in one place. The prompt stays on your device. Currency can display in USD or INR. That choice is not a copy of your text.
LLM price check for one prompt
The token calculator is the first step in an AI model cost comparison. Paste the system message, user turn, or excerpt you would send on one call. The page counts characters and words as you type, then prices one model per family. GPT tokens use tiktoken in this tab. Claude, Gemini, and Grok counts are labeled estimates, not those vendors’ tokenizers.
The chart is a one-call LLM price for the prompt. Each bar shows that model’s cost and token count. If the number is high, trim repeated instructions, drop unused context, or switch to a cheaper model in the same family. That is the LLM cost optimization this site can do: see the price, then cut the prompt. When you open the monthly calculator from that result, input tokens carry over from the GPT count. Output starts at zero because there is no reply yet. Type an expected answer length if you want completion tokens in the bill.
LLM cost calculator for a month of calls
The API cost page is a separate LLM cost calculator for traffic, not a second paste box. Enter input and output tokens per request, requests per day, users, and days in the month. The same usage is priced for one Claude, GPT, Gemini, and Grok model so you get an LLM cost comparison as a monthly sketch. Direct visits start empty. Arriving from the token page fills the prompt’s token count. Use 1 user if requests per day is already a team total. Leave output at 0 to bill the prompt only. The figure is for planning, not an invoice and not a live quote from OpenAI, Anthropic, Google, or xAI.
AI model price comparison and catalog rates
AI model price moves. This site does not scrape a live ticker on every view. Catalog input and output rates per million tokens sit on the rates page, taken from the OpenRouter public list for the allowlisted models. That list is the source for the LLM price comparison in the charts. OpenAI model price for the GPT rows follows that catalog, as do the other families. A spreadsheet of dollars per million tokens does not tell you what your text costs. The token page does. A single-call LLM price does not tell you what a launch costs. The monthly calculator does, once you add traffic.
FinOps, attribution, and what this site is not
Some teams want to track AI model cost attribution in FinOps stacks. Online Model Cost does not connect to billing exports, tag traces, or allocate spend to services. It is a calculator. You can still take a number into the system that does attribution: this prompt is this many tokens; this many requests per day is this monthly LLM cost. Prompt-layer LLM cost optimization is what this page can do. Warehouse-level allocation is not.
For a catalog LLM price comparison without a prompt, opencatalog rates. For an LLM price calculator on one call, open the token calculator. For a monthly AI model cost comparison, open theAPI cost calculator. All three stay in the browser. None of them upload the prompt.