Free Token Counters and LLM Cost Calculators Compared: og.cc.cd, OpenAI Tokenizer, Tiktokenizer, tiktoken and Vendor APIs
TL;DR: For a quick private count plus a cost comparison across 33 models, use the og.cc.cd token counter; for exact OpenAI tokenization use OpenAI’s Tokenizer, Tiktokenizer or the tiktoken library; for billing-grade Claude or Gemini counts use the vendor’s free token-counting API.
Free token counters answer different questions. Some show exactly how OpenAI’s tokenizer splits text, some build a full chat payload, and the vendors’ own APIs give billing-grade counts for their models. Here is what each common free tool does, from its own page, and where our tools fit.
Open the Token Counter →All guides
Comparison table
| Tool | Made by | What it counts | Where it runs | Cost comparison | Best for |
|---|---|---|---|---|---|
| og.cc.cd Token Counter | og.cc.cd | Exact o200k_base and cl100k_base (js-tiktoken); other vendors labelled as estimates | In your browser, no upload, no sign-up | 33 models, 13 vendors, official prices with sources | A private first count and “what would this cost on each model?” |
| OpenAI Tokenizer | OpenAI | OpenAI tokenization | Web page on platform.openai.com | — | Checking OpenAI tokenization from the source |
| Tiktokenizer | dqbd (open source) | Chat messages for a chosen model (e.g. gpt-4o), with whitespace view | Web app | — | Seeing how a multi-message chat payload is counted |
| tiktoken | OpenAI (MIT licence) | Any text with o200k_base, cl100k_base and other encodings | Python library (pip install tiktoken) | — | Counting in code, batches and CI |
| Anthropic count_tokens | Anthropic | A full Claude request incl. system prompt, tools, images and PDFs | API (free, rate-limited; needs a key) | — | Exact Claude counts before you send |
| Gemini countTokens | Input of a Gemini request, including images, video and audio | API / SDK (needs a key) | — | Exact Gemini input counts | |
| OpenAI input token counting | OpenAI | A full Responses API payload (POST /v1/responses/input_tokens) | API (needs a key) | — | Exact OpenAI counts including tools and images |
When our tools are the right choice
- You want a count and a price in one step. The token counter shows the o200k and cl100k counts and a cost table for every model we track, each price linked to the vendor page.
- You are budgeting a workload, not a single prompt. The cost calculator turns requests per day, average tokens and cache-hit share into a monthly bill, cheapest first.
- The text is private. Both tools run in your browser with no account and no upload, and they keep working offline once the page has loaded.
When to use something else
- You need billing-exact counts for Claude or Gemini. We only estimate non-OpenAI models (Claude ×1.3 on the o200k count, because Anthropic says Claude 4.7 and later produce about 30% more tokens). Use Anthropic’s free count_tokens endpoint or Gemini’s countTokens instead.
- You count in code or CI. Use the tiktoken library (or the vendors’ APIs) so the count runs where your prompts are built.
- You are debugging a multi-message chat payload. Tiktokenizer’s message builder, or the vendor counting endpoints, include the per-message formatting tokens.
Whatever you use before sending, the usage field in each API response is the final word on what you were billed.
Free LLM cost calculators
| Option | What it gives you | Limits |
|---|---|---|
| og.cc.cd LLM Cost Calculator | Daily and monthly cost for 33 models from requests, tokens and cache share; long-prompt tiers applied | List prices only: no batch discounts, cache-write fees, taxes or free tiers |
| Vendor pricing pages | The authoritative price for that vendor | One vendor at a time; you do the maths |
Your API usage dashboard and the usage field | What you actually spent | Only after the fact |
FAQ
What is the best free token counter?
For a private count plus a cost comparison across many models, the og.cc.cd token counter; for exact OpenAI tokenization, OpenAI’s Tokenizer, Tiktokenizer or the tiktoken library; for exact Claude or Gemini counts, the vendor’s token-counting API.
Is the og.cc.cd token counter accurate?
The o200k_base and cl100k_base counts are exact (computed with js-tiktoken). Rows for Claude, Gemini and other vendors are clearly labelled estimates, because those vendors do not publish a browser tokenizer.
Do free token counters upload my text?
Ours does not: counting runs in your browser. Vendor counting APIs receive the request you send them, like any API call. Check each third-party site’s own privacy policy.
Is Anthropic’s token counting API free?
Anthropic’s docs say token counting is free to use but subject to requests-per-minute limits based on your usage tier, and that the count is an estimate that may differ slightly from actual usage.
Sources
Vendor features and prices were checked against these official pages on 2026-10-22. Products change; if something here no longer matches, the official page is right.
- OpenAI — tiktoken (GitHub)
https://github.com/openai/tiktoken - OpenAI — Tokenizer
https://platform.openai.com/tokenizer - OpenAI — Counting tokens
https://developers.openai.com/api/docs/guides/token-counting - Tiktokenizer
https://tiktokenizer.vercel.app/ - Tiktokenizer — source code
https://github.com/dqbd/tiktokenizer - Anthropic — Token counting
https://docs.anthropic.com/en/docs/build-with-claude/token-counting - Gemini API — Understand and count tokens
https://ai.google.dev/gemini-api/docs/tokens