Ping Model

Measure AI model response latency with consistent formatting.

Quick Start

Simple Ping (current model)

Ping Specific Model

Compare Multiple Models

Command Reference

Output Format

Required format - ALWAYS use this exact structure:

CODEBLOCK3

Latency Formatting Rules

• - < 1 second: Display as XXXms (e.g., 847ms)
• ≥ 1 second, < 60 seconds: Display as X.XXs (e.g., 1.23s)
• ≥ 60 seconds: Display as X.XXmin (e.g., 2.5min)

Examples

Fast response (< 1s):
CODEBLOCK4

Medium response (1-60s):
CODEBLOCK5

Slow response (> 60s):
CODEBLOCK6

Cross-Model Testing

When testing a non-active model:

1. 1. Save current model context
2. Switch to target model
3. Execute ping
4. Measure latency
5. Restore original model
6. Display result

Critical: Always return to the original model after testing.

Comparison Mode

CODEBLOCK7

Output format:
CODEBLOCK8

Implementation

The ping latency is measured as the time between:

• - T1: Message received by the agent
• T2: Response ready to send

This captures the model's internal processing time, not network latency.