Skip to main content

Documentation Index

Fetch the complete documentation index at: https://koreai.mintlify.app/llms.txt

Use this file to discover all available pages before exploring further.

Back to Analytics Overview The Performance Analytics page provides comprehensive insights into LLM performance across models, features, and requests, enabling data-driven decisions for optimization and enhanced user experience. To access the report, go to Analytics > Gen AI Analytics > Performance Analytics. You can also access this report from the dashboard by selecting a model or feature in its corresponding widget. Performance Analytics Apply filters at the top of the page by module, model, feature, and date range to focus on specific interactions and drill down into targeted performance metrics.

​
Field Descriptions

You can sort the data alphabetically or by values in ascending or descending order. Click any record to view detailed model performance metrics, including request trends, token consumption patterns, and latency analysis. Performance Analytics Details
FieldDescription
ModelThe Large Language Model to which the request was made.
FeatureThe platform feature making calls to the LLM models.
RequestsTotal number of LLM calls made for that model-feature combination.
Successful RequestsNumber of LLM calls that executed without errors.
Failed RequestsNumber of LLM calls that encountered errors. Together with successful requests, provides an execution breakdown to monitor system reliability.
Total TokensCombined token count for request and response.
Request TokensThe individual parts of input text (words, punctuation) given to the model to create a response. These tokens form the basis for the model’s understanding.
Response TokensThe pieces of generated output (words, punctuation) showing the model’s response. These tokens form the structured output of the model.
Avg. Tokens per RequestAverage token consumption for each request.
Avg. Tokens per ResponseAverage token consumption for each response.
Average Response TimeThe mean time taken by the model to return a response.
Median LatencyMedian processing time, useful for understanding response consistency.