Tema_Q-X4-Thinking (Tema Q) is an improved large-scale language model (LLM) for Japanese and English based on the model Gemma 4 26B A4B developed by Google.
It is designed to generate more flexible and useful responses to prompts that are difficult to answer using the standard Gemma model.
Note: It is believed that inference accuracy will drop significantly below Q2_K. If possible, please use Q3_K_M.