DeepSeek API tests show Claude-like behavior under selected prompts
The results revive Anthropic's distillation claims, though the test cannot establish who generated the outputs or whether they were used for training.
By Ryan Merket · Published
Primary source: X - @synthwavedd
Why it matters
DeepSeek released V4 before Fable 5 existed, so any verified overlap would point to post-release data collection, live routing or an undisclosed API update rather than V4's original training.
An independent test of DeepSeek-V4-Pro found sharply different behavior depending on whether prompts contained cyber or biology content, raising new questions about whether some requests were being answered with outputs from Anthropic's Claude Fable 5.…