TokenCALC
AI token & cost calculator
Free privacy-first AI token calculator. Count tokens and estimate LLM API cost in one browser tool across OpenAI, Anthropic, Google, DeepSeek, and more. Text stays on your device.
Tokenizer · Cost calculator · Providers · Compare · Tools · Guides
Tools
Small utilities next to the main calculator. Each one answers a single job.
Cost calculator
Paste a prompt, tune cache and batch, compare models, and share the estimate URL. Token text never leaves this browser.
- Input tokens
- 65
- Output tokens
- 512
- Input cost
- $0.00013
- Output cost
- $0.004096
- Characters
- 306
- Words
- 51
Monthly projector
- Daily
- $2.113
- Monthly
- $63.39
- Yearly
- $771.25
Rates in $2.00/1M · out $8.00/1M
Context 577 / 1,000,000 · Verified · official pricing
gpt-tokenizer · o200k_base
Token pieces
ExactCompare models
Same prompt, output size, cache, and batch settings across models.
| Model | Accuracy | Input tok | Request | Monthly |
|---|---|---|---|---|
DeepSeek V4 FlashDeepSeek | Approx | 77 | $0.000154 | $2.31 |
Gemini 2.5 FlashGoogle | Approx | 77 | $0.001303 | $19.55 |
GPT-4.1OpenAI | Exact | 65 | $0.004226 | $63.39 |
Claude Sonnet 5Anthropic | Approx | 77 | $0.005274 | $79.11 |
Pricing catalog
120 curated models · 10 providers · prices checked against official pages (not scraped)
Official pricing sources
How it works
- Paste a prompt, document, or chat message.
- Pick a model. Exact for OpenAI encodings in your browser; Approx when the provider tokenizer is not available client-side.
- See token count and API cost. Optionally set cache hit %, Batch, and monthly volume for a budget, not just a per-request number.
What is a token? · How LLM pricing works · Prompt cost · Exact vs Approx
Deeper reading: tokenization · tokens vs words · OpenAI · Claude · Gemini · prompt caching · context windows · all guides
Compare popular models
Same pricing catalog as the calculator. Open a pair, then run your own prompt. Showing 12 highlights of 102 featured comparisons.
- GPT-5 vs Claude Opus 4.8
- GPT-5 vs Gemini 3.1 Pro
- Claude Opus 4.8 vs Gemini 3.1 Pro
- GPT-4.1 vs Claude Sonnet 5
- GPT-4.1 vs Gemini 2.5 Pro
- Claude Sonnet 5 vs Gemini 2.5 Pro
- GPT-5 Mini vs Claude Sonnet 4.6
- GPT-4.1 Mini vs Claude Haiku 4.5
- GPT-4o mini vs Gemini 2.5 Flash
- Gemini 2.5 Flash vs Claude Haiku 4.5
- Gemini 3 Flash vs GPT-5 Mini
- GPT-4.1 Nano vs Gemini 2.5 Flash-Lite
Privacy & trust
Built as a privacy-first cost workstation: accuracy labels and local processing are product features, not footnotes.
Text stays local
Prompts are tokenized in your browser. We do not upload your text to a server.
Honest Exact / Approx
OpenAI encodings are Exact via gpt-tokenizer. Other providers are labeled Approx, never silently faked.
Curated prices
Rates are manually verified against official pages with lastVerified dates and outbound source links. No scrape theater.
Share without a backend
Use Copy share link when you want a bookmarkable URL with your estimate. Day to day browsing keeps clean addresses.
FAQ
- Does TokenCALC send my text to a server?
- No. Tokenization and cost estimates run in your browser. Your prompt text is not uploaded to TokenCALC servers.
- What do Exact and Approx mean?
- Exact uses a matching OpenAI tokenizer encoding in the browser. Approx means the provider tokenizer is not available in-browser, so we use a labeled heuristic instead of faking precision.
- Are the API prices live?
- Prices are curated from official provider docs (with a weekly sync script + PR review). Each model has a lastVerified date and link to the official page. We do not pretend rates are a live streaming API.
- Can I project monthly LLM spend?
- Yes. Set active users and messages per user per day to project daily, monthly, and yearly cost from your per-request estimate, including cache and batch when published.