Users are reporting that Anthropic's Claude models, including Opus 4.8 and Fable, have become excessively filtered in recent weeks. Specific instances include the model refusing to discuss open-source firmware for consumer electronics and providing only basic election statistics when asked about their impact on government contracts. This increased filtering is causing frustration among paying users who rely on the service for coding and other tasks. AI
IMPACT Increased filtering may impact user trust and adoption of Claude models for sensitive or complex queries.
RANK_REASON User discussion about model behavior and filtering.
AI-generated summary · Google Gemini · from 1 sources. How we write summaries →