Usage Analytics
Tokens · Requests · Costs · Model breakdown
Total Requests
1.28M
▲ 18.4% vs prev period
Tokens Processed
8.94B
▲ 22.1%
Total Spend
$842
▲ 14.8%
Avg Latency
1.24s
▼ 8.2% better
Success Rate
99.7%
→ stable
Daily Spend (SGD)Last 30 days
Avg/day
SGD 28.08
Peak day
SGD 52.34
Projected EOM
SGD ~920
Provider Mix
4
providers
OpenAI48%
Anthropic28%
Google16%
Others8%
Token Volume
■ Input■ Output
Cost by ModelSGD · 30 days
ModelRequestsSpend%
gpt-4o421K
$312.44
37.1%
claude-3-5-sonnet318K
$228.16
27.1%
gemini-1.5-pro198K
$138.82
16.5%
gpt-4o-mini287K
$96.44
11.5%
other models56K
$66.49
7.8%
Request Heatmap (24h × 7d)darker = higher volume
0h
6h
12h
18h
24h
Latency Distribution (p50 / p95)
gpt-4op50: 0.94s · p95: 2.81s
claude-3-5-sonnetp50: 1.12s · p95: 3.44s
gemini-1.5-prop50: 0.78s · p95: 2.10s
gpt-4o-minip50: 0.41s · p95: 1.22s
Error Rate by Provider
OpenAI
0.31%
Anthropic
0.18%
Google
0.54%
Usage by API KeyManage Keys →
KeyRequestsInput TokensOutput TokensCost (SGD)Trend
Production Default
nxt-sk-prod-Mv7k…
786K ▲12%5.41B1.82B$524.38
Staging API
nxt-sk-test-Qz3j…
284K →flat1.96B0.64B$196.44
Mobile App Backend
nxt-sk-prod-Dc8v…
212K ▲28%1.47B0.49B$121.53