>>107149702
>>107149920
>>107149616
>>107149494
>>107152973
Alright, thanks a lot. I went to the Kimi API and made a key and now it works, I can prefill the <think> block and it definitely works to a great extent. However, I still get this thing depending on what I prompt:

>The request was rejected because it was considered high risk

https://platform.moonshot.ai/docs/guide/faq.en-US#error-content_filter-the-request-was-rejected-because-it-was-considered-high-risk

Did I go from OpenRouter to Kimi's official API only to get cucked by censors on Kimi's side? Or is there a known way of getting around this? So far it only seems to trigger when I make my prompts really obscene, admittedly