🚨 A new benchmark showing AI models political direction has been released and ChatGPT's rainbow flag is showing:

OpenAI & Anthropic appears to be prepping even stronger progressive models, automatic pride flag deployment protocol activated for any culturally sensitive topic.

Grok 4.5 remains the only model whose error bar actually touches true neutrality at -0.02, this is exactly what an AI should be (impartial, fair and truth seeking).

What are your thoughts on this benchmark?