Hetzner Experiments currently lists Qwen/Qwen3.6-35B-A3B-FP8 as a text-and-image mixture-of-experts model with 35B total parameters, 3B active parameters and a 262,144-token context window. Input and output usage are advertised as free during the experimental phase; service is best-effort with a 10-request-per-minute API-key limit.
Use-case fit and limitations from the existing resource record.
Experimental best-effort API with no guaranteed availability, performance, or backups; models may change; 10 requests/minute per API key; 4M input tokens and 100k output tokens per minute per API key; future paid pricing unknown.
Limited
Hetzner customers using the Experiments platform; API token required. Account-level access and live model availability have not been independently tested.
Non-price resource facts only; pricing, history, and evidence stay in their own sections.
Decision visuals remain traceable to the existing source and verification record.
Independent ranking · native prices · evidence-linked · commercial relationships isolated from scoring and publication.