Hetzner explores LLM inference with experimental API for testing purposes

Jul 24, 2026

AI Summary

Hetzner is conducting experiments with a new API for large language model (LLM) inference, currently featuring a single model, Qwen/Qwen3.6-35B-A3B-FP8. The initiative aims to assess user interest and system capabilities without any production guarantees or billing.

Hetzner is testing an LLM inference API that is compatible with OpenAI's standards.
The current model available for testing is Qwen/Qwen3.6-35B-A3B-FP8, which has 35 billion parameters and accepts both text and images.
The API is in an experimental phase, lacking billing, service level agreements (SLAs), and production guarantees.
Users can create an API token through the Experiments dashboard and integrate it with their applications using standard OpenAI client libraries.
The initiative is aimed at understanding user demand, system scalability, and feature requirements.
Hetzner's approach may leverage its efficient hardware acquisition and operational capabilities to potentially enter the low-margin inference market.
The current GPU offerings from Hetzner may not support larger models, raising questions about future hardware investments for scaling the service.
The experiment is seen as a way to utilize spare GPU capacity and generate revenue, but its long-term viability depends on further developments in hardware and model offerings.

llminferencehetznerai researchtechnology

Hetzner explores LLM inference with experimental API for testing purposes

Related Stories

Grok 4.6 Achieves 61 on Artificial Analysis Intelligence Index, Enhancing SpaceXAI's Position

Exploring the Mathematical Capabilities of Large Language Models

Google's Gemini app reaches over 1 billion monthly active users

Anthropic AI model advances understanding of the Riemann hypothesis