chiprook
← AI
AIOctober 25, 2026, 11:35

Moonshot curbs hacking ability in Kimi K3.1 Pro open-weight model

Moonshot AI released Kimi K3.1 Pro, a 3-trillion-parameter-class open-weight model trained without cyber-related data, with weights due by November 4. It scored 81.0% on the company's internal cyber benchmark versus 90.5% for a version without the restrictions.

Moonshot curbs hacking ability in Kimi K3.1 Pro open-weight model
#Moonshot#Kimi
Read next
AI

Kimi K3.1 Pro launch faces scrutiny over undisclosed influencer promotion

AI

Leak: Moonshot preps Kimi K3.1 launch before October

AI

First Proof: Open-Weight AI Trails Frontier Models by Ten Months in Math

AI

Moonshot releases open-source Kimi K2.7-Code with lower token usage