induwara.lk
induwara.lkAI · Pricing

AI Cost Per Word Calculator — ChatGPT, Claude & Gemini

Token pricing is awkward when you bill clients by the article. This tool converts GPT-4o, Claude, and Gemini list prices into cost per word, per page, and per article — in dollars and rupees — and ranks every model cheapest-first. No signup, sources cited below.

By Induwara AshinsanaUpdated Jun 9, 2026
Cost per word
List prices · verified 2026-06-09
What are you doing?

The AI writes the words — billed at the model's output price.

How many words you'll generate or process.

Quick sizes

CBSL indicative; override freely.

How many of these you make per month.

Advanced — tokens per word1.333 tok/word

0.5–5. Default 1.333 (English).

1,000-word generate (output) on GPT-4o
Rs 4.00$0.0133 · 1,333 tokens
Per word
Rs 0.003999
$0.00001333
Per 100 words
Rs 0.40
$0.001333
Per page (500 w)
Rs 2.00
$0.006665
Per article (1,000 w)
Rs 4.00
$0.0133

Model comparison

ModelUSDLKRvs cheapest
Gemini 1.5 FlashCheapest
Google · $0.3/1M for this task
$0.0003999Rs 0.12—
GPT-4o mini
OpenAI · $0.6/1M for this task
$0.0007998Rs 0.24+100%
Claude 3.5 Haiku
Anthropic · $4/1M for this task
$0.005332Rs 1.60+1233%
Gemini 1.5 Pro
Google · $5/1M for this task
$0.006665Rs 2.00+1567%
GPT-4o
OpenAI · $10/1M for this task
$0.0133Rs 4.00+3233%
Claude 3.5 Sonnet
Anthropic · $15/1M for this task
$0.02Rs 6.00+4900%

Estimates from list prices and an averaged token↔word ratio. Real token counts vary by tokenizer, model, and language.

How it works

Large-language-model APIs bill in tokens, not words — and they charge separately for the tokens you send in and the tokens the model writes out. For a content writer that is the wrong unit. This calculator does the conversion deterministically, entirely in your browser, so you can see what a draft actually costs.

  1. Words → tokens. Multiply the word count by the tokens-per-word ratio (default 1.333, the reciprocal of OpenAI's “~0.75 words per token”). Non-English or code-heavy text uses more tokens per word, so the ratio is adjustable.
  2. Pick the rate. Generate uses the model's output price; Process uses the input price; Both adds the two (read N words, write N words). Output is dearer than input on every model here because the model spends compute generating each token.
  3. Tokens → USD. List prices are quoted per million tokens, so costUSD = tokens ÷ 1,000,000 × pricePerMillion. The per-word figure is this divided by the word count, then scaled ×100, ×500, and ×1,000 for the per-100-words, per-page, and per-article cards.
  4. USD → LKR. Multiply by the USD→LKR rate (default 300, a Central Bank of Sri Lanka indicative figure you can override). A monthly total multiplies the per-piece cost by your volume.
  5. Comparison.Steps 2–4 repeat for all six preset models, sorted cheapest-first for your exact word count and task type, with each model's premium over the cheapest shown as a percentage.

Every cost is cross-checked against an independent per-thousand-token formula (tokens ÷ 1,000 × rate ÷ 1,000) that must agree with the per-million path to the fraction of a cent — the same belt-and-braces approach the tax calculator uses against the IRD's formula. All figures are list prices and an averaged ratio; for exact token counts, use the linked token counter.

Worked examples

Generate a 1,000-word article on GPT-4o

Generate · output $10/1M · 1.333 tok/word · Rs 300/USD

  1. tokens = 1,000 × 1.333 = 1,333
  2. costUSD = 1,333 ÷ 1,000,000 × 10 = $0.01333
  3. per word = $0.01333 ÷ 1,000 = $0.0000133 (Rs 0.004)
  4. per page (500 w) = $0.0000133 × 500 = $0.0067
  5. in LKR: $0.01333 × 300 = Rs 4.00 per article
  6. 40 articles/month: $0.01333 × 40 = $0.533 ≈ Rs 160

Same article on Gemini 1.5 Flash, compared

Generate · output $0.30/1M · 1.333 tok/word · Rs 300/USD

  1. costUSD = 1,333 ÷ 1,000,000 × 0.30 = $0.0004 (Rs 0.12)
  2. ≈ 33× cheaper than GPT-4o for the same words
  3. Cheapest-first: Gemini 1.5 Flash $0.0004
  4. < GPT-4o mini $0.0008 < Claude 3.5 Haiku $0.0053
  5. < Gemini 1.5 Pro $0.0067 < GPT-4o $0.0133
  6. < Claude 3.5 Sonnet $0.0200

Edge case — process a 5,000-word document on Gemini 1.5 Flash

Process · input $0.075/1M · 1.333 tok/word · Rs 300/USD

  1. tokens = 5,000 × 1.333 = 6,665
  2. costUSD = 6,665 ÷ 1,000,000 × 0.075 = $0.0004999
  3. cross-check: 6.665 thousand-token units × $0.000075 = $0.0005 ✓
  4. in LKR: $0.0005 × 300 = Rs 0.15
  5. Reading uses the input price, so it's far cheaper than writing.

Frequently asked questions

Sources & references

The model prices and token ratio on this page were last cross-checked against the official sources on 2026-06-09. AI pricing changes often — if a figure looks off, email me and I'll re-verify within 24 hours.

Related tools

LiveAI

AI Agent Cost Calculator

Estimate the real per-run, daily, and monthly cost of a multi-step LLM agent across Claude, GPT, and Gemini. Models the context accumulation single-call calculators miss — each tool result is re-sent every step — with caching, in USD and LKR.

Open tool
LiveAI

AI Chatbot Cost Calculator

Estimate the monthly API cost of a multi-turn AI chatbot across Claude, GPT, and Gemini. Models the quadratic context re-sending that single-call calculators miss, with and without prompt caching, in USD and LKR.

Open tool
LiveAI

AI Rate Limit Calculator

Computes whether an LLM workload will hit OpenAI, Anthropic, or Gemini rate limits — effective max requests/min, which limit binds (RPM vs TPM/ITPM/OTPM/RPD), and batch wall-clock time.

Open tool
Rate this tool
Be the first to rate

Comments & feedback

Spotted a bug or want an improvement? Tell us — our team reviews every comment, and good ideas get built. Comments are public and anonymous.

Spotted a pricing change, edge case, or want another model added?

Email me at [email protected] — most fixes ship within 24 hours.