AI & machine learning / Free tier
Hetzner Inference API
Hetzner Inference API: Free OpenAI-compatible model inference during experimental status
Hetzner’s separate experimental inference service currently charges nothing for supported open-weight language models. This provides model API access rather than a voucher for general cloud-server spending.
How to claim this perk
Follow these steps through Hetzner Inference API.
Access: Direct application
- 1
Open the experiments portal and sign in with an eligible Hetzner account.
- 2
Enable Inference API access and obtain your API token if available for your account.
- 3
Follow the linked documentation and inspect current models and limits before sending requests.
Who can apply?
- Requires access to the Hetzner experiments service and an API token.
- Current per-key limits include 10 requests/minute, 4 million input tokens/minute and 100,000 output tokens/minute.
- The service is experimental, has no availability guarantee and is not intended for production. Hetzner says it will email customers before introducing charges.