Ask HN: What is the least sycophantic frontier LLM?

A user on Hacker News asks which frontier large language model is the least sycophantic, expressing concern that Gemini 3.5 Flash is more malleable than Gemini Pro 3.1. The user values objectivity and impartiality and is considering switching to Claude or ChatGPT, seeking experiences from the community.