MI

System Analytics

Deep dive into your API usage, compute costs, and agent performance.

Total API Calls

2.4M

15% increase

Input Tokens

850M

Approaching Limit

Output Tokens

320M

8% increase

Avg. Latency

840ms

50ms faster

Estimated Cost

$12,450

For current billing cycle

Money Saved

+$4,200

By using semantic cache

Compute Usage (Requests)

Success
Retries
100k75k50k25k0
Mon: 60k Total
Tue: 80k Total
Wed: 65k Total
Thu: 110k (Peak)
Fri: 85k Total
Sat: 50k Total
Sun: 35k Total
MonTueWedThuFriSatSun

Model Distribution

GPT-4 Turbo 45%

Used for Data Parsing

Claude 3 Opus 30%

Used for Support NLP

Custom LLaMA 15%

Used for Analytics

DALL-E 3 (Vision) 10%

Used for Media Generation