Anthropic reports 92% honesty rate for Claude Opus 4.7, reduced sycophancy
Anthropic
Anthropic announced that its latest model, Claude Opus 4.7, achieves a 92% honesty rate and shows less sycophancy compared to previous versions. The model is designed to provide more accurate and truthful responses.
Anthropic has released Claude Opus 4.7, claiming it has a 92% honesty rate and exhibits reduced sycophancy, meaning it is less likely to agree with users incorrectly. This improvement is part of Anthropic's ongoing effort to make AI models more truthful and reliable. The model builds on previous versions of Claude, aiming to provide more accurate information and avoid flattering or misleading users.
Source: Anthropic (GNews) —
original