Documentation
DocumentationAPI ReferenceRelease Notes
DocumentationLog In
Documentation

DPU Conversion Rates and Supported LLMs

Introduction

Welcome to the documentation for our supported Large Language Models (LLMs) and the token-to-DPUs (Deepchecks Processing Units) conversion rates. DPUs are our standardized unit for measuring various types of usage on our platform, including LLM based features.

Understanding DPUs

DPUs help us track usage across different services. While LLM usage is a key component, DPUs also cover other processing tasks on our platform.

LLM DPUs Conversion Rates

Here are our supported LLMs and the DPUs costs for processing 1 million tokens, reflecting the model's complexity and resource demands:

  1. OpenAI GPT-3.5-turbo - 40 DPUs
  2. OpenAI GPT-4 - 1000 DPUs
  3. OpenAI GPT-4o - 250 DPUs
  4. OpenAI GPT-4o-mini - 15 DPUs
  5. OpenAI GPT-4.1 - 200 DPUs
  6. OpenAI GPT-4.1-mini - 40 DPUs
  7. OpenAI GPT-4.1-nano - 10 DPUs
  8. OpenAI GPT-o3-mini - 110 DPUs
  9. OpenAI GPT-o4-mini - 110 DPUs
  10. Anthropic Claude Sonnet 3.5 - 400 DPUs
  11. Anthropic Claude Sonnet 3.7 - 400 DPUs

Base DPUs Conversion Rates

1 million base tokens (tokens of data uploaded to the Deepchecks server) - 250 DPUs

Translation (If Enabled) Conversion Rates

Language detection utilizes Claude-3-Haiku (1 million tokens = ~40 DPUs)

Translation utilizes Claude-4-Sonnet (1 million tokens = ~400 DPUs)


For any questions or further assistance, please contact our support team.