KHATT — Benchmark & Integration

Real API Latency Benchmark & Setup for Cursor, 9router & n8n

Honest Time-to-First-Token (TTFT) numbers measured directly from our production gateway, plus how to connect your favorite coding and automation tools in under a minute.

1. Real Measured Latency (Streaming TTFT)

Instead of quoting synthetic ping numbers, here are real streaming chat/completions measurements (stream: true, max_tokens: 16) from our production node:

Model Streaming TTFT Total Time Sample Cost
gemini-3.8-flash ~1.02s ~1.08s ~0.07 cr/reply
gpt-6.1-sol ~1.13s ~1.34s ~0.20 cr/reply
deepseek-v4.1-flash ~1.0s ~1.1s 0.05 cr (min floor)
claude-opus-5.5 ~5.05s ~5.07s ~0.33 cr/reply

You can verify gateway health and endpoint responsiveness yourself anytime via https://api.khatthub.com/health or curl:

curl -s -o /dev/null -w "ttfb:%{time_starttransfer}s http:%{http_code}\n" https://api.khatthub.com/v1/models

2. Connect Cursor, 9router, or n8n in 30 Seconds

  1. Open @KhattAIbot on Telegram and run /key to generate your API key (includes 3 free trial credits).
  2. In Cursor, 9router, or n8n (OpenAI Chat Model node), set:
    • Base URL: https://api.khatthub.com/v1
    • API Key: your sk-khatt-... key
    • Model: any of the 12 curated models (e.g. gpt-5.5, claude-opus-5, deepseek-v4-flash)
  3. Every request automatically runs through KHATT's GCC PII Shield and auto-fallback routing with pay-as-you-go credits.

Test real latency and model routing with 3 free trial credits — pay-as-you-go from AED 39 (~50% less than ChatGPT Plus).

Get API Key in @KhattAIbot →

© 2026 KHATT. ← All guides · AED Cost Calculator · PII Proof · Home