Dev Tools · 1h ago
Hetzner Tests LLM Inference with Experimental API
Hetzner launched an experimental, OpenAI-compatible inference API offering a single 35B-parameter MoE model. The service is free, has no SLA, and is intended to gauge user interest and system scalability. Developers can try it via a simple API key and standard OpenAI client.
Meridian48 take
A low-risk experiment from a major European hosting provider, but without pricing or production guarantees, it's more a curiosity than a viable alternative to established inference APIs.
hetzner-inferencellm-api