Real-time micro-probe benchmark measuring Time-to-First-Token (TTFT), throughput (TPS), and true Token pricing across 13+ premier models (DeepSeek, Qwen 2.5, Claude, GPT-4o) from edge gateways. Built with transparent micro-probes and an interactive monthly cost simulator.

How did you use Cloud Run for this launch?

We leverage Google Cloud Run to deploy containerized micro-telemetry probe workers that run automated, sub-second latency (TTFT) and throughput (TPS) benchmarks against 13+ premier LLM models. Cloud Run's serverless architecture and container elasticity allow us to execute scheduled probe jobs with zero idle compute costs, providing independent edge telemetry and real-time model comparison data to developers globally.

Hey Product Hunt community! 👋 We built SkyBreak Platform because we were tired of relying on synthetic marketing claims rather than real-world connection speeds. For developers worldwide, actual Time-to-First-Token (TTFT) and cold-start latency vary drastically across providers. Key highlights of SkyBreak: ⚡ Real Edge Micro-Probes: Tests 13+ premier models every 30 mins from edge gateways with sub-second TTFT and TPS measurement. 📊 True Token Economics: Tracks live pricing per 1M prompt/completion tokens with Prompt Caching discounts. 🧮 Monthly Cost Simulator: Estimate your real monthly production bill across all models in seconds. 💻 Drop-in Code Snippets: 1-click Python, Node.js, and cURL snippets compatible with the OpenAI SDK. We’d love to hear your feedback! Which models or metrics would you like to see benchmarked next?

How did you use Cloud Run for this launch?

SkyBreak Platform

Live latency & pricing benchmark for 13+ global & Asian LLMs

Live latency & pricing benchmark for 13+ global & Asian LLMs