Mistral Exposed for Distilling DeepSeek and Faking Results
Mistral accused of distilling DeepSeek's model & faking benchmarks. Whistleblower exposes alleged misconduct. Open-source AI giant faces controversy.
"AI Disruption" Publication 7400 Subscriptions 20% Discount Offer Link.
Resignation Drama! Mistral Exposed for "Distilling" DeepSeek
A Twitter user broke the news that a female former Mistral employee sent mass emails exposing multiple company scandals.
The most explosive revelation: Mistral's latest model allegedly directly distilled from DeepSeek, yet packaged it externally as a successful RL case while deliberately distorting benchmark test results.
Speaking of Mistral, this company is hailed as the European version of OpenAI and is one of the global open-source star players, with model performance consistently receiving high praise.
Precisely because of its outstanding reputation, this exposé appears particularly shocking.
