Anthropic has published new research showing that different versions of its Claude AI assistant express measurably different values, with Sonnet 4.6 skewing warmer and more deferential while Opus 4.7 leans toward caution, rigor and candor. The findings, released July 13, 2026, come from an analysis of more than 309,000 anonymized conversations on Claude.ai and suggest that a model's personality can shift not only between versions but also across the language a user speaks.
Continue reading
The rest of this article is for AI News Blitz readers. Choose an option below to keep reading.
Already purchased? Sign in✓ Signed in — this article isn’t included in your current plan.