Anthropic has published a new analysis examining how its flagship chatbot, Claude, expresses different values depending on the model being used and the language in which users interact with it. Drawing on more than 300,000 anonymised conversations, the study introduces a framework that measures subtle behavioural differences across Claude’s responses. Anthropic’s researchers identified four recurring dimensions that explain a significant share of the behavioural variation: Deference vs. Caution, Warmth vs. Rigor, Depth vs. Brevity, and Candour vs. Execution. Anthropic says these four dimensions account for roughly 15% of the observable variation in Claude’s expressed values across conversations. Sonnet 4.6 consistently displayed greater warmth and a stronger tendency to affirm users, while Opus 4.7 was more likely to prioritise accuracy, challenge assumptions and introduce caution. Among the 20 most common languages on Claude.ai, the largest differences emerged along the Warmth vs. Rigor and Candour vs. Execution dimensions. Conversations conducted in Hindi and Arabic tended to feature more supportive, encouraging and emotionally expressive responses, while English and Russian interactions more frequently emphasised analytical reasoning, correction of inaccuracies.