Nova

Model & Provider Health

Compare route reliability, response performance, and cost across your gateway

openai
GPT-5.5 Pro
OpenAI

Route success

100.0%

P95 latency

1.5s

P50 TTFT

320ms

Gateway overhead

39ms

Fallback rate

15.6%

Cache served

6.3%

40.1K requests$271 / 1K req
anthropic
Claude Opus 4.7
Anthropic

Route success

95.8%

P95 latency

1.5s

P50 TTFT

362ms

Gateway overhead

37ms

Fallback rate

12.5%

Cache served

16.7%

89K requests$33 / 1K req
google
Gemini 3.1 Pro
Google

Route success

100.0%

P95 latency

806ms

P50 TTFT

154ms

Gateway overhead

41ms

Fallback rate

7.1%

Cache served

25.0%

68.6K requests$18 / 1K req

Routing signals

Best current candidates across the selected model pool

Most reliable

GPT-5.5 Pro

100.0% route success

Fastest response

Gemini 3.1 Pro

154ms median TTFT

Lowest route cost

Gemini 3.1 Pro

$18 / 1K requests

Provider Route Health

Live attempt reliability across primary and fallback routes

5 providers active
openaiOpenAI
Monitor

Success

90.8%

Rate limit

6.1%

P50 TTFT

264ms

P95

1.1s

anthropicAnthropic
Monitor

Success

84.5%

Rate limit

8.7%

P50 TTFT

258ms

P95

1.3s

googleGoogle
Monitor

Success

85.7%

Rate limit

8.6%

P50 TTFT

152ms

P95

637ms

mistralMistral
Operational

Success

100.0%

Rate limit

0.0%

P50 TTFT

165ms

P95

640ms

deepseekDeepSeek
Degraded

Success

40.0%

Rate limit

40.0%

P50 TTFT

305ms

P95

1.0s

Latency Distribution by Model
Daily Cost by Model
Gateway Performance Comparison

Usage totals paired with route-level performance for the selected models

ModelProviderRequestsSpendSuccessP50 TTFTP95 latencyCacheFallbackCost / 1K req
openaiGPT-5.5 Pro
OpenAI40.1K$8,838100.0%320ms1.5s6.3%15.6%$271
anthropicClaude Opus 4.7
Anthropic89K$2,82695.8%362ms1.5s16.7%12.5%$33
googleGemini 3.1 Pro
Google68.6K$1,021100.0%154ms806ms25.0%7.1%$18