AI Translation Cost Calculator

Cheap enough to translate everything — almost

LLM translation collapses the cost of going multilingual, but proofread anything high-stakes. For bulk jobs, batch pricing cuts it further — see the batch savings calculator.

Machine translation: API vs LLM pricing

Dedicated translation APIs (Google, DeepL, Amazon) price per character — about $10–$25 per million characters. LLMs restructured this market: translating via a cheap model costs $1–$5 per million characters equivalent at current token prices, with quality that matches or beats dedicated APIs on major language pairs. The catch is throughput and glossary control — translation APIs offer terminology enforcement, format preservation, and batch endpoints that LLM pipelines rebuild by hand. Under 500k characters monthly, free tiers cover everything; above that, the LLM route usually wins on price if you can tolerate building the plumbing.

How this calculator works

The AI Translation Cost Calculator estimates what it costs to translate a body of text with a large language model. It takes your word count, the number of target languages, and the model's input and output prices per million tokens, then multiplies them out into a total. The main cost drivers are volume and language count: each additional target language repeats the full translation, so cost scales roughly with words times languages. Because models bill separately for the text you send and the text they generate, both the input and output rates matter, and output is often priced higher than input.

The key trade-off to watch is that translated output frequently runs longer than the source, so the output side of the bill can dominate — especially into verbose languages or when you translate into many locales at once. A practical tip: check the price before scaling up, since a rate that looks trivial per thousand words can grow quickly across large documents and multiple languages. Use the calculator to compare models and confirm that a cheaper input rate is not offset by a higher output rate for your specific workload.

Frequently asked questions

How much does AI translation cost?

Roughly the source words in and the translated words out, each ~1.33 tokens/word. Cost = (input tokens × input price + output tokens × output price) per language. A small model makes bulk translation very cheap.

Is LLM translation good enough?

For most content it's excellent and far cheaper than human translation. For legal, medical or brand-critical copy, have a human review — the cost of an error outweighs the savings.