Community discusses rising costs of AI hardware for local models

A Reddit user reports ~23 tok/s on Gemma 4 26b with an RTX 4060 Ti 16GB, highlighting performance limitations. The discussion thread (83 comments) focuses on the financial burden of upgrading hardware for local AI models.