Open API PricingOpen API Pricing

Google Cloud Translation API vs DeepL API pricing

Both are in translation apis. Google Cloud Translation API is billed per million chars; DeepL API is billed per million chars. Numbers below are copied from each vendor's pricing page.

Google Cloud Translation API vs DeepL API pricing side by side
Free tier
500k chars/mo (never expires)
Billed
Per million chars
Best for
Hyperscaler MT at global scale
Verified
Jun 27, 2026
DeepL APIFree tier or trial
Free tier
500k chars/mo (API Free)
Billed
Per million chars
Best for
High-quality MT API for developers
Verified
Jun 27, 2026

Google Cloud Translation API plans

PlanPriceWhat it covers
Free tier$0First 500,000 characters/month free (credit applies collectively across Basic and Advanced editions)
NMT text (Basic v2 / Advanced v3)$20 / 1M charsStandard neural machine translation over the free tier; Basic over 1B chars/month is contact-sales
NMT document translation (Advanced)$0.08 / pageDOCX/PPTX/PDF; max 4,000 chars or 500 words per page
Translation LLM$10 in + $10 out / 1M charsLLM-based text translation, priced separately on input and output characters
Adaptive Translation (LLM)$25 in + $25 out / 1M charsLearns from supplied examples to adapt to style/domain
Custom AutoML models$80 / 1M chars (tiered down to $30)$80/M under 250M, $60/M to 2.5B, $40/M to 4B, $30/M above; training $45/hr capped at $300/job

Source: https://cloud.google.com/translate/pricing

DeepL API plans

PlanPriceWhat it covers
API Free$0/monthUp to 500,000 characters per month; full text and document translation API access, no cost.
API Growth~$26/month (annual)Includes ~12M characters/year (~1M/month) with overage billed around $27.50 per additional million characters.
Enterprise APICustomUp-front custom commitments for characters, speech-to-text and speech-to-speech; pay-as-you-go above commitment; includes 99.9% uptime SLA.

Source: https://www.eesel.ai/blog/deepl-pricing

Google Cloud Translation API pricing guideDeepL API pricing guideQuality scores on APIbenchmarks →