Estimate recurring charges for partner-delivered services.

NMT stands for Neural Machine Translation, Google's machine learning-based translation model for natural text translation.

If you do not specify a source language for the translate method, Cloud Translation detects the source language for you. You are only charged for the text that you provide; there is no additional charge for language detection on top of translation.

Reference: Google Cloud Translation pricing

* Price is per character sent to the API for processing. For details about counted characters, see Charged characters.

This model reflects event-driven Cloud Run request, CPU, and memory pricing only. It does not include always-on warm instances, jobs, GPUs, or networking-linked charges.

Singapore (`asia-southeast1`) is used as the baseline region. The calculator intentionally excludes idle baseline behavior so the comparison stays closer to AWS Lambda and Azure Functions Consumption.

The free tier is applied directly as free vCPU-seconds, free GiB-seconds, and free requests.

Reference: Google Cloud Run pricing

Amazon Translate charges by processed character count, including whitespace.

Singapore is the baseline account region for this calculator. Amazon Translate pricing is generally service-level rather than strongly region-differentiated, but taxes and account context can still vary.

Reference: Amazon Translate pricing

* The AWS free tier for Amazon Translate is time-limited for new usage. Set this to `0` if you want a conservative steady-state estimate.

This model reflects Lambda request and duration pricing only. It does not include provisioned concurrency, response streaming, or extra ephemeral storage charges.

This calculator uses standard AWS Lambda on-demand baseline pricing in USD. Validate final regional and commercial pricing against the actual AWS account that will run the workload.

Reference: AWS Lambda pricing

Azure AI Translator pricing is shown by Microsoft as a consumption-based estimate and actual pricing can vary by agreement and billing context.

Singapore is the baseline for this calculator. Translator is generally treated as a non-regional resource, but account pricing and currency treatment can still vary.

Reference: Azure AI Translator pricing

This default uses GPT-5 mini pricing as a lightweight text-processing baseline. You can overwrite the model label and token rates if you plan to use a different OpenAI model.

OpenAI pricing is token-based rather than character-based, so this estimate should be treated separately from the Google, AWS, and Azure translation calculators for now.

As a rough guide for English text, 1 token is about 4 characters, so 1 million characters is often around 250,000 tokens. Tokenization varies by language, punctuation, and formatting, and non-English text can use more tokens per character.

Cached input pricing is optional. Leave cached share at 0% if you expect each translation request to be unique.

Reference: OpenAI API pricing

This model reflects Azure Functions request and execution-time pricing only. It does not include premium plans, networking, or storage-linked charges.

Singapore is the baseline billing context for this calculator. The default rates here are aligned to the Azure Functions Consumption plan, not Flex Consumption.

Reference: Azure Functions pricing

Uses S3 Standard pricing only. No lifecycle, infrequent access, or intelligent tiering assumptions are included.

Uses Standard object storage only. No archive tiers, replication premiums, or lifecycle transitions are included.

Uses Blob Storage Hot tier only. No cool/archive tiers, redundancy uplifts, or lifecycle rules are included.