Users question alleged routing of Kimi and DeepSeek queries to Claude
Two users argue that Kimi and DeepSeek’s immediate reasoning traces would reveal a switch to Claude, which they say would not provide those traces in real time. One asks Anthropic for clarification.
TLDR
One user describes Anthropic’s allegation as Chinese entities redirecting Kimi and DeepSeek users to Claude, then separately recovering reasoning traces through an alleged weakness in Anthropic’s API. Two users challenge that scenario on timing: they say Kimi and DeepSeek display reasoning traces—step-by-step reasoning output—immediately, whereas the alleged Claude routing approach would not. They argue that this mismatch would expose the switch, with one calling on Anthropic to clarify its account.
Combined views
18.6K
2 Sources, first seen 20d ago