What is LLM Overwatch?

LLM Overwatch is a tool that monitors the response times, performance metrics, and operational status of AI providers (OpenAI, Anthropic, and Gemini) and their models.
It includes model-specific breakdowns, historical trends, peak/off-peak analysis, and incident reports.

Note: The response times are measured by generating a maximum of 512 tokens with a randomized prompt every 10 minutes in 3 locations. The maximum response time is capped at 60 seconds but could be higher in reality. Percentage changes (↑ or ↓) represent the relative difference compared to the all-time average, while the values in parentheses show the absolute response time difference from that average.

Last data update: --:--:--()
Next update in: --:--

OpenAI

Loading...

Last 15 min. average response time:

--

Anthropic

Loading...

Last 15 min. average response time:

--

Gemini

Loading...

Last 15 min. average response time:

--

Peak/Off-Peak Hours Response Times

All Providers
OpenAI
Anthropic
Gemini

All Providers Hourly Avg. Response Times

Note: This chart shows the average response times across all providers for each hour of the day. Peak hours (higher response times) are shown in purple, while off-peak hours (lower response times) are shown in green.

OpenAI Hourly Response Times

Anthropic Hourly Response Times

Gemini Hourly Response Times

Model Response Time Comparison

All Models
OpenAI Models
Anthropic Models
Gemini Models

All Models Hourly Comparison

Note: This chart shows the hourly response times for all models across providers. Each line represents a different model, with colors indicating the provider. The data helps identify performance patterns and compare models across different times of day.

OpenAI Models Hourly Comparison

All-Time Average Response Times by Model

Loading model data...

Anthropic Models Hourly Comparison

All-Time Average Response Times by Model

Loading model data...

Gemini Models Hourly Comparison

All-Time Average Response Times by Model

Loading model data...

Recent Incidents

All
OpenAI
Anthropic
Gemini

Loading incident data...