PANews April 24 news, according to Anthropic blog, Anthropic recently admitted that Claude Code has experienced a decline in code quality, due to three product and prompt changes: first, on March 4, the default reasoning effort was lowered from high to medium, resulting in a decline in intelligence, which was restored to high/xhigh on April 7; second, on March 26, a bug in the clear_thinking_20251015 logic introduced for optimizing caching caused long idle times to continuously lose reasoning history, leading to forgetting and repetition, fixed on April 10 in v2.1.101; third, on April 16, the addition of the “≤25/100 word” system prompt, included to compress word count, had about 3% negative impact on code quality, which was rolled back on April 20. Anthropic stated that the API inference layer was not affected, and has reset usage quotas for all subscribers.

免责声明：本文章仅代表作者个人观点，不代表本平台的立场和观点。本文章仅供信息分享，不构成对任何人的任何投资建议。用户与作者之间的任何争议，与本平台无关。如网页中刊载的文章或图片涉及侵权，请提供相关的权利证明和身份证明发送邮件到support@aicoin.com，本平台相关工作人员将会进行核查。