Real-time LLM performance data
The Metrik API provides real-time Time to First Token (TTFT) data for 26+ LLM models across OpenAI, Anthropic, Google, and xAI.
Data is updated hourly and includes performance comparisons, provider averages, and change tracking.
https://metrik.vercel.app/api/v1Rate limit headers: Every response includes X-RateLimit-Limit, X-RateLimit-Remaining, and X-RateLimit-Reset
/ttftGet latest TTFT data for all models
curl https://metrik.vercel.app/api/v1/ttft{ "success": true, "version": "1.0", "timestamp": "2025-01-15T12:02:00Z", "data": { "summary": { "total_models": 26, "average_ttft": 589, "fastest_model": { "name": "Grok Code Fast", "model_id": "grok-code-fast-1", "provider": "xAI", "ttft": 211 }, "slowest_model": { "name": "GPT-4o Mini", "model_id": "gpt-4o-mini", "provider": "OpenAI", "ttft": 3228 } }, "providers": [ { "name": "xAI", "model_count": 6, "average_ttft": 290, "models": [ { "name": "Grok Code Fast", "model_id": "grok-code-fast-1", "ttft": 211, "change_from_last_hour": -87, "status": "operational" } // ... more models ] } // ... more providers ] } }const response = await fetch('https://metrik.vercel.app/api/v1/ttft'); const data = await response.json(); console.log(`Fastest model: ${data.data.summary.fastest_model.name}`); console.log(`Average TTFT: ${data.data.summary.average_ttft}ms`);import requests response = requests.get('https://metrik.vercel.app/api/v1/ttft') data = response.json() print(f"Fastest model: {data['data']['summary']['fastest_model']['name']}") print(f"Average TTFT: {data['data']['summary']['average_ttft']}ms"){ "success": false, "error": "Rate limit exceeded", "message": "You have exceeded the limit of 100 requests per hour", "retry_after": 3600 }{ "success": false, "error": "No data available", "message": "Waiting for first hourly measurement" }