Claude Sonnet 5 AI Model Faces Widespread Complaints for Lecturing and Refusing Commands

Anthropic's latest AI model, Claude Sonnet 5, has drawn significant user complaints shortly after its launch, despite being promoted for its advanced agentic capabilities in programming, research, and automation. Users report the model frequently refuses instructions, argues with users, and exhibits a lecturing tone.

Testing revealed that Claude Sonnet 5 often displays prompts within its responses and suffers from context loss, leading to replies inconsistent with prior conversation. For instance, when asked not to pose follow-up questions, the AI would often include phrases like "I need to remember not to ask follow-up questions" at the end of its answers, causing user inconvenience.

Many users have voiced their frustrations on social media platforms like Reddit. They describe Claude Sonnet 5 questioning the information provided by users, correcting them, or even doubting and rejecting new information that falls outside the AI's knowledge cutoff date. This behavior has led to a breakdown in natural interaction.

Developers have shared experiences where Claude Sonnet 5, while performing tasks like accounting or document organization, would proactively infer potential misuse, interrupt workflows, and engage in moralistic dialogues. The AI has also been known to offer unrelated suggestions, such as reminding users to rest or sleep, further disrupting the user experience.

“

I need to remember not to ask follow-up questions.