All AI calculators
Showing 8 of 8 AI calculators
AI Context Window Calculator
Calculate how much context your prompt, conversation and documents need, and whether they fit in the model’s context window.
Open toolAI Cache Savings Calculator
Work out how much prompt caching cuts your AI input bill, including cache-write costs and cache misses.
Open toolAI Batch Cost Calculator
Price a large AI job through a batch API and see what it saves over sending the same requests in real time.
Open toolAI Automation Cost Calculator
Add up everything an AI workflow really costs each month, not just the tokens, and see whether the time it saves pays for it.
Open toolAI Inference Cost Calculator
Turn token counts and per-million prices into cost per request, per month and per year for an LLM API.
Open toolAI OCR Cost Calculator
Work out what OCR plus AI extraction costs per page, per document and per month, including the people who check it.
Open toolAI GPU Requirement Calculator
Estimate how much GPU memory a model needs for inference or fine-tuning, and which class of GPU that points to.
Open toolAI Image VRAM Calculator
Estimate how much GPU memory an image generation workflow needs, and what it takes to run it on a smaller card.
Open toolBuilding with AI models raises the same few questions again and again. What will this cost once real users arrive? Would caching or batching make it cheaper? Will the prompt fit? Can this GPU run that model? The answers are mostly arithmetic, but arithmetic with a lot of moving parts - input and output prices, cache writes and reads, bytes per parameter, KV cache per token.
These calculators do that arithmetic and show their working, so you can see which assumption drives the result. They take prices, token counts and model sizes as inputs rather than quoting them as fixed facts, because all three change constantly.
Which AI calculator do you need?
- To budget an LLM feature per request, per month and per year, use the AI Inference Cost Calculator.
- To see how much reusing a long system prompt or shared documents saves, use the AI Cache Savings Calculator - it accounts for cache writes and misses, not just the discount.
- To price a large offline job such as classifying a backlog, use the AI Batch Cost Calculator.
- To cost a whole workflow or agent - model calls, tools, platform and human review - and its ROI, use the AI Automation Cost Calculator.
- To cost document processing per page and per document, and compare it with manual data entry, use the AI OCR Cost Calculator.
- To check that a prompt, its history and documents fit alongside the answer, use the AI Context Window Calculator.
- To estimate GPU memory for LLM inference or LoRA, QLoRA and full fine-tuning, use the AI GPU Requirement Calculator.
- To estimate VRAM for Stable Diffusion, SDXL or Flux-class image generation, use the AI Image VRAM Calculator.
Why prices and sizes are inputs, not facts
Model prices fall, new models appear every few months, and providers change how caching and batching are billed. A calculator that hard-codes today’s prices is quietly wrong a few months later. Here, every price is a field you can edit; where a preset is offered, it carries the date it was checked and where it came from.
The same goes for hardware estimates. GPU memory use depends on the model’s architecture and on the software running it, so the GPU and image VRAM calculators give an estimate, a recommendation with headroom and a table of common card sizes - not a single “correct” GPU.
Where the money usually goes
- Output tokens cost several times more than input on most APIs, so long answers dominate many bills.
- Long, repeated prompts are where caching helps - the context window and cache savings calculators show how big that repeated part is.
- In automations, human review time often costs more than the model itself.
- For self-hosting, memory - not raw compute - usually decides which GPU a model needs.
Frequently asked questions
Are these AI calculators free?
Yes. No sign-up, no limits and no premium tier.
Do I have to enter my prompts or data?
No. You enter numbers - token counts, prices, volumes, model sizes - not your prompts or documents, and the calculations run in your browser.
How current are the prices?
Where a calculator offers provider presets, each one says which pricing page it came from and the date it was checked. Prices change often, so check the provider’s page before relying on a figure. You can always type your own.
How accurate are the GPU and VRAM estimates?
They are planning estimates built from standard formulas - bytes per parameter, KV cache per token, optimizer state - plus overhead and headroom. Real usage varies with the framework, settings and model architecture, so measure on your own setup before buying hardware.
How do I count tokens?
The most reliable source is the usage data your provider returns with each API response, or its token-counting tool. For rough English text, a token is about four characters or three quarters of a word.
Other free tools on VSThiran
Free OCR tool
Extract text from images and scanned PDFs on your own device.
Free developer tools
JSON, Base64, hashes and strong passwords.
Free text tools
Word and character counts, case conversion and whitespace cleanup.
Free PDF tools
Convert, merge, split and compress PDFs without uploading them.