I Added Three Rules to Gemma 4. The MoE Searched. The Dense Model Refused.

TL;DR: I run an AI sales chatbot for Arabic-speaking merchants. I wanted to know if Gemma 4 could replace GPT-4o-mini on the customer-facing reply. I tested two Gemma 4 variants — the 26B mixture-of-experts (4B active params) and the 31B dense model — against GPT-4o-mini and GPT-4o, across six Arabic customer scenarios, through my real production chat router. The actual failure mode of both Gemma variants in Round 1 wasn't hallucination. It was reluctance — stalling instead of searching, hedging...