In the ongoing rivalry between Gemini 2.5 Pro and Claude 3.7 in the AI coding assistant landscape, various users have shared feedback based on practical coding experiences. While Gemini 2.5 Pro is praised for its powerful reasoning abilities and impressive context management, users have pointed out several areas where it falls short. The comments reveal that Gemini occasionally generates unnecessary refactorings and fails to follow explicit instructions well, particularly in complex codebases. On the other hand, Claude 3.7 appears to have weaknesses in adapting to new coding styles but shows strengths in following user instructions and code improvement tasks. Many users express disappointment with both models, observing that while minor improvements continue, they are not substantial enough to justify the model's costs compared to free alternatives. The consensus suggests a growing concern about the ability of established models to maintain their competitive edge against innovative, open-source solutions.