Elastic Inference Service (EIS) cost estimation

There are two separate charges for using Elastic Inference Service (EIS):

• Elastic Managed LLM:
  • $4.50 per million input tokens
  • $21.00 per million output tokens
• Elastic Inference Service (EIS): As low as $0.08 per million input tokens

Note: EIS is currently not available on StackOps Elastic, so you will only be billed for using the Elastic Managed LLM.

Estimating tokens

A rough token rule of thumb is ~750 words ≈ 1,000 tokens. In a typical AI Assistant exchange, assuming ~500-word prompt + ~300-word response, the cost would roughly be (667 input tokens × $4.50/M) + (400 output tokens × $21/M) = $0.003 to $0.01 per exchange.

Example:

At scale, if a customer has 10,000 AI Assistant interactions/month with an average of 1,000 input tokens + 500 output tokens each interaction:

• Input cost: 10M tokens × $4.50/M = $45
• Output cost: 5M tokens × $21/M = $105
• Total cost: ~$150/month for that volume