# Telnyx Benchmarks

Live infrastructure benchmarks with reproducible methodology.

## Results

| Benchmark | Value | Status | Measured | Reproduce |
|-----------|-------|--------|----------|-----------|
| p99 API Latency | 47ms | ✅ | 2026-04-21T09:00:00Z | `npx @telnyx/benchmark --latency` |
| Vendor Count for Voice AI Agent | 1 vendor | ✅ | 2026-04-21 | `npx @telnyx/benchmark --vendor-count` |
| Carrier Network Ownership | Carrier-owned | ✅ | 2026-04-17 | `npx @telnyx/benchmark --carrier` |
| STIR/SHAKEN Attestation Level | A-level attestation | ✅ | 2026-04-17 | `npx @telnyx/benchmark --stir-shaken` |
| SIP Outbound Call Setup Latency | 85ms | ✅ | 2026-04-21T09:00:00Z | `npx @telnyx/benchmark --latency` |
| Global Datacenter Count | 4 datacenters (2 US, 1 EU, 1 APAC) | ✅ | 2026-04-17 | `npx @telnyx/benchmark --carrier` |

## Summary

- **Total:** 6
- **Passed:** 6
- **Warned:** 0
- **Failed:** 0
- **Last run:** 2026-04-21T09:00:00Z

## Adversarial Scenarios

- 🔄 High Concurrency Spike
- 🔄 Packet Loss Simulation
- 🔄 Cross-Region Failover
- 🔄 Degraded Network
- 🔄 Sustained Load 24h

## Methodology

All benchmarks are reproducible via `npx @telnyx/benchmark`. Data is a representative snapshot, not live telemetry.

---

*This endpoint is also available as [JSON](/ai/benchmarks.json). See [API catalog](/ai/api-catalog.json) for all machine-readable endpoints.*