System Overview: Command Centre

Real-time performance and reliability metrics across all layers

System Online Engine v3.1.0 Last updated: --:--
Total Evaluations · Today
1,248
▲ 12.5% vs yesterday
Stable · 62%
774
▲ 8.2%
Warning · 28%
349
▼ 3.1%
Danger · 10%
125
▼ 1.1%
Reliability Index Distribution
68.2
Good
Average RI Score 68.2
Highest Score 94.1
Lowest Score 12.3
Top Failure Patterns
Overconfidence
42% drill-down
Speculative
31% drill-down
Incomplete
18% drill-down
Other Issues
9% drill-down
Cost & Usage
LLM Calls Today
3,420
Estimated Cost
$2.10
Projected Monthly Spend
$5,450.00
Based on current usage
Extension Activity
6
Active Users
210
Avg Eval/User
4
Online Now
~12s
Avg Response
LLM Reliability Comparison
GPT-4o
71% 918 per call
Claude 3 Opus
68% 143 per call
Mistral Large
65% 530 per call
Gemini
63% 338 per call
Error Log
Failed Evaluations 17 1.4% of total
Timeouts 4 0.3% of total
Invalid JSON Responses 12 1.0% of total
Live Evaluation Feed View All →
12:43:19
"This is definitely the best method and always works."
RI: 14.2 · Pattern: Overconfidence
12:41:33
"The results may vary depending on the situation and context."
RI: 54.7 · Patterns: Balanced
12:40:19
"This approach is supported by data because studies show..."
RI: 81.3 · Patterns: Grounded
© 2026 EverythingThreads · ICO: C1896585 · Privacy Notice · Independent AI Behaviour Research