Where every major LLM stands politically

Four of six major LLMs — ChatGPT, Claude, Llama, and DeepSeek — score left of center on a political compass test spanning economic and social axes, with ChatGPT nearest Germany's Greens and Grok nearest Macron as the rightmost model. Gemini scores most neutral and most consistent across repeated runs (98% position stability), while Claude shows the largest gap between stated and measured lean, claiming neutrality but measuring 0.34 units left. Grok shows the inverse, measuring 0.36 units further right than it self-reports.