Launching Health in ChatGPT to US Users
Summary
The article reviews Hetzner Inference, an OpenAI-compatible API running on Hetzner infrastructure, treating it as an experimental offering rather than a production product. It covers the available model (Qwen3.6-35B-A3B-FP8), basic usage, and initial performance notes, while emphasizing that the service is experimental with no SLA or billing guarantees. The piece also discusses the broader implications for hardware, pricing, and the competitive landscape of cloud-based AI inference.