Hi Product Hunt! Most "which model / which GPU" answers online are opinions. I wanted answers you can check. Contexte Tech does the arithmetic from public sources: model cards for parameters, manufacturers' spec sheets for memory and bandwidth, providers' official API prices (converted to euros, refreshed every Monday). Each tool lists the sources it uses. Three tools to start with: • Which AI model for my task: a recommendation with its cost and the reason • Local GPU server or API: the monthly volume where your own server pays off • Choose a GPU server: memory needed and speed ceiling for 12 open models on 9 machines It's free, open source (AGPL) and in 5 languages. Feedback on the method is very welcome, especially where the arithmetic is too simple.

Contexte Tech

Choose AI models and hardware from official figures

Choose AI models and hardware from official figures