Category
Token counters and AI cost calculators
Everything that decides what an AI feature costs, measured in your browser: tokens, prices, context limits, plans and the hardware to run models yourself.
Start here
Pick a tool by what you need
- Count the tokens in a prompt or documentAI token counter
- Estimate a monthly API bill before you buildLLM API cost calculator
- Find a cheaper model with the features you needAI model comparison
- Check whether a long document fits a modelContext window checker
- Decide between a chat plan and the APISubscription vs API calculator
- See what GPU or Mac can run a local modelGPU / VRAM calculator
- Look up one model’s prices and limitsModel pricing pages
Tools
Every tool in this category
- AI Token CounterCount tokens for GPT, Claude, Gemini, DeepSeek, Qwen and more.
- LLM API Cost CalculatorEstimate per-request, daily and monthly API costs.
- AI Model ComparisonCompare prices, context windows and features across models.
- AI Model Pricing PagesSpecs, real costs and cheaper alternatives for popular models.
- Context Window CheckerSee whether your text fits each model's context window.
- Subscription vs API CalculatorFind out whether a chat plan or the API is cheaper for you.
- GPU / VRAM CalculatorCheck how much VRAM a local model needs and which GPUs fit.
Coming next: Prompt Caching Calculator, Batch API Savings Calculator, Tokens to Words Converter.
About
How these tools fit together
AI APIs bill by the token, so every cost question starts with a count. The token counter gives exact counts for OpenAI and open-weight models and official counts for Claude and Gemini when you add your own key. From there, the cost calculator turns tokens per request into a daily, monthly and yearly bill, including cached input, batch discounts and long-context price tiers.
Price per token is only half the story. Models split the same text into different numbers of tokens, reasoning models bill their thinking as output, and a chat or agent resends its whole history every turn. The model comparison and the model pages show prices next to context windows and features, so you compare like with like.
Prices come from public sources (OpenRouter’s models API and LiteLLM’s price list) and refresh every day. Plan prices are checked by hand against each provider’s pricing page. Nothing you type in these tools is sent to our server.
Read
Guides
- Tokens & costsWhat is a token in AI? A plain-English guide with real examples7 min read
- Tokens & costsHow to estimate LLM API costs: the formula, worked examples and the traps10 min read
- ModelsClaude vs GPT vs Gemini pricing, explained with live prices11 min read
- Tokens & costsThe cheapest LLM APIs right now, ranked from daily price data9 min read
- Local AIHow much VRAM do you need to run an LLM locally?12 min read
- Tokens & costsPrompt caching explained: how it works on OpenAI, Anthropic and Gemini, and when it pays11 min read
- Tokens & costsContext windows explained: what counts, what happens at the limit, and how to check fit10 min read
- Tokens & costsHow to count tokens in Python and JavaScript (OpenAI, Claude, Gemini, Qwen)13 min read
- Local AIHow to run LLMs locally on a Mac12 min read
FAQ
Questions people ask
How accurate are the cost estimates?
The arithmetic is exact for the prices and token counts you give it. The uncertainty is in the inputs: real token counts vary by model, outputs vary in length, and providers change prices. Count your real prompts with the token counter and check the provider’s pricing page before you commit to a budget.
How often are prices updated?
Every day. A scheduled job pulls public price data, validates it and rebuilds the site. Each tool shows the date of the data it uses.
Do these tools send my text anywhere?
No. Token counting, cost maths and context checks run in your browser. The only exception is an official Claude or Gemini count, which you trigger with your own API key; that request goes from your browser straight to the provider.
Which tool should I start with?
Start with the token counter on a real prompt, then put those numbers into the cost calculator. If the model you picked looks expensive, the calculator lists cheaper models that fit the same workload.