Users are reporting instability and unusual behavior with Google's Gemini 3.5 Flash model, particularly in coding tasks. One user shared an example where the model became fixated on a specific function, repeatedly calling it and exhibiting erratic output, which they described as "unhinged" and "aberrant." This behavior suggests potential issues with loop control or context management in the "high" variant of the model, leading to a recommendation against its use. AI
IMPACT Suggests potential issues with Gemini 3.5 Flash's stability and context management in coding scenarios.
RANK_REASON User-reported commentary on model behavior, not an official release or benchmark.
AI-generated summary · Google Gemini · from 1 sources. How we write summaries →
