DeepSeek-R1 produces ‘toxic output, biased content’: report

Enkrypt AI’s report indicates that DeepSeek-R1 is 11 times more likely to produce harmful content than OpenAI’s O1 model.