Local language models · model catalogue
gpt-oss-120b for private hosting in Europe
Efficiency model. An open model optimised for cost and speed: ideal for high volumes, classification, extraction and everything that needs to scale.
0.15 €
Input · reference price € / 1M tokens
0.60 €
Output · reference price € / 1M tokens
128K
Context · Tokens
2
Location options: Berlin · Vienna
At mixed usage it is 80× cheaper than Claude Fable 5 at its API list price ($10 / $50 per 1M tokens, USD taken 1:1).
Operation and access by agreement
We plan operation of the model on-premise or at a suitable European location. Before activation we confirm availability, API address, model identifier and terms. Data retention, logging, subprocessors and compliance requirements are agreed for your configuration.
Enquire about your model and location without obligation. We confirm capacity and the start date individually.