Machine translation pricing compared
Translation APIs and LLM alternatives, priced by characters or tokens. Test your language pairs and terminology.
13 services · sources last checked 2026-10-06 to 2026-10-09 · how we compare
What the most-compared services cost
Estimated yearly bill at 50M characters / month. Enter your own usage below.
Enter your usage to estimate each bill. Services differ in what they include; check the rates, assumptions and extra charges below.
The estimates below use the default usage. Choose your current provider to compare costs.
| Provider | Parent jurisdiction | Features / pricing source | Est. monthly | Est. yearly |
|---|---|---|---|---|
| GPT-6 Luna (via OpenAI API)Translation with OpenAI's small GPT-6 Luna model, priced from OpenAI API token rates rather than per character. | 🇺🇸 United States | EU residencyGDPR DPAModeled ratesToken-to-character conversion assumes English source, short prompts and European-language output. Quality and retries are not priced. | $9.25$0.185 per M chars | $111 |
| Qwen3.8-Flash (via Alibaba Cloud Model Studio)Translation with Alibaba's small Qwen3.8-Flash model on Model Studio (international), priced from token rates. | 🇨🇳 China | GDPR DPAModeled ratesToken-to-character conversion assumes English source, short prompts and European-language output. Quality and retries are not priced. | $9.68$0.1935 per M chars | $116 |
| Mistral Small 4 (via Mistral API)Translation with Mistral's open-weight Small 4 model on Mistral's EU-hosted La Plateforme API, priced from token rates. | 🇫🇷 France | EU residencyGDPR DPAModeled ratesToken-to-character conversion assumes English source, short prompts and European-language output. Quality and retries are not priced. | $12$0.2325 per M chars | $140 |
| DeepSeek V4 Flash (via DeepSeek API)Translation with DeepSeek's open-weight V4 Flash model on DeepSeek's first-party API in China, priced from token rates. | 🇨🇳 China | Modeled ratesToken-to-character conversion assumes English source, short prompts and European-language output. Quality and retries are not priced. | $23$0.465 per M chars | $279 |
| Gemini 3.1 Flash-Lite (via Gemini API)Translation with Google's small Gemini 3.1 Flash-Lite model, priced from Gemini API token rates rather than per character. | 🇺🇸 United States | GDPR DPAModeled ratesToken-to-character conversion assumes English source, short prompts and European-language output. Quality and retries are not priced. | $27$0.5375 per M chars | $323 |
| Azure AI TranslatorMicrosoft's Translator API (part of Azure AI Foundry Tools) with custom models, document translation and monthly commitment tiers. | 🇺🇸 United States | EU residencyGDPR DPAPublished rates | $500S1 pay-as-you-go · $10 per M chars | $6,000 |
| Amazon TranslateAWS neural machine translation API with real-time and batch document translation, custom terminology and Active Custom Translation. | 🇺🇸 United States | EU residencyGDPR DPAPublished rates | $750$15 per M chars | $9,000 |
| ModernMTAdaptive neural MT API from Italy's Translated that learns from translation memories and corrections; being replaced by Lara, sunsets end of 2026. | 🇮🇹 Italy | Published rates | $750$15 per M chars | $9,000 |
| SYSTRAN Translate APINeural MT API from French veteran SYSTRAN (ChapsVision group), hosted in Europe, with specialised industry models and user dictionaries. | 🇫🇷 France | EU residencyPublished rates | $840$16.79 per M chars | $10,074 |
| Google Cloud TranslationGoogle's Cloud Translation API (NMT) with batch and document translation, glossaries, custom AutoML models and an LLM-based Translation tier. | 🇺🇸 United States | EU residencyGDPR DPAPublished rates | $990$20 per M chars | $11,880 |
| Lara TranslateLLM-based translation API from Italy's Translated (successor to ModernMT) with context, style and translation-memory adaptation. | 🇮🇹 Italy | GDPR DPAPublished rates | $1,029Team · $19.99 per M chars | $12,354 |
| DeepL APIGerman neural MT API with glossaries, formality control and document translation; EU-hosted, billed per character on top of a small base fee. | 🇩🇪 Germany | EU residencyGDPR DPAPublished rates | $1,238API Pro · $24.64 per M chars | $14,851 |
LLM rows are computed from token prices (assumptions in notes). Glossaries, document formats and quality vary, so test on your own content.
Need EU residency? Turn on the EU filter to pick a plan that includes it, which may cost more. “EU plan available” alone does not mean the price shown includes residency. Shared links save your inputs and filters, but prices can change.
Looking for a European-owned provider? See European machine translation providers.
Self-hosted options
Operating cost not estimated · servers, licenses and engineering time are yours. We list these separately because there’s no calculated bill to compare with managed services.
- LibreTranslate — AGPL-licensed, self-hostable machine translation API built on Argos Translate (OpenNMT/CTranslate2); paid hosted keys also offered.
No self-hosted options match these filters.
Head-to-head comparisons
Machine translation pricing explained
What drives the bill
Translation APIs usually charge per million characters. Small LLMs charge by tokens and can cost less on some workloads. This calculator converts token prices to a per-character estimate, assuming English source text, prompt overhead and a certain output length. These rows are labeled as modeled rates. Check each provider’s notes and test quality in the language pairs you use.
Costs the calculator leaves out
- Document translation (DOCX, PDF), often priced per page or at a higher per-character rate.
- Glossaries and terminology control, which some APIs support natively and which take prompt work with LLMs.
- Quality review. For published content, human review usually costs far more than the API.
- Low-resource languages, where quality varies most between engines.
When paying more makes sense
Specialist translation APIs give consistent output, formality controls, glossary support and predictable behavior without prompt engineering, which matters in production pipelines. For user-generated content, support tickets and internal documents, a small LLM may be worth testing. Compare on a few hundred real sentences in your language pairs, including terminology and failure cases, before you decide.