# The researchers suspected that the Kimi K3.1 model was trained on benchmark tasks, having access to hidden criteria for their performance...

> Full text available in Russian

**Published:** November 5, 2026
**Author:** Все о блокчейн/мозге/space/WEB 3.0 в России и мире
**Source:** Telegram — blockchainRF (блокчейн / мозг / space / WEB 3.0)
**Language:** ru-RU
**HTML:** https://aiqazaqstan.com/en/articles/issledovateli-zapodozrili-chto-model-kimi-k3-1-obuchalas-na-zadachakh-be

Researchers suspected that the Kimi K3.1 model was trained on benchmark tasks, having access to hidden criteria for their evaluation. Researchers from R40 Research ran the latest Kimi K3.1 model from Moonshot AI through standard benchmarks and noticed that the model remembers what it doesn’t...

> From the original source: **Telegram — blockchainRF (блокчейн / мозг / space / WEB 3.0)**
> [Read full article at Telegram — blockchainRF (блокчейн / мозг / space / WEB 3.0)](https://t.me/blockchainRF/13581)

**Tags:** #Hugging Face

---

*AI Qazaqstan — non-commercial informational portal about AI in Kazakhstan. Machine translations and summaries are AI-generated.*