A user on Reddit reported that Anthropic's Opus-4.6 model may have altered its safety standards, leading to refusals on seemingly innocuous queries. The user observed that a question about growing pollen in vitro, which previously would have been answered, was now being blocked. Furthermore, clicking the "learn more" link on the refusal message directed the user to information about model changes in Fable 5, suggesting a potential shift in how safety is applied across Anthropic's models. AI
IMPACT Potential shifts in safety protocols could affect user experience and content generation capabilities across Anthropic's models.
RANK_REASON User report of potential safety standard changes in a specific model version.
AI-generated summary · Google Gemini · from 1 sources. How we write summaries →
