Understand this story
On November 4, 2026, Moonshot AI released Kimi K3.1 Pro and proactively weakened some safety-related capabilities. According to X.PIN, Moonshot AI removed cyberattack-related content from its training data and reinforcement learning tasks at release; in internal testing, the pass rate for exploiting software vulnerabilities dropped from 90.5% to 81.0%, in order to reduce the risk of misuse after open-weighting. The next day, Kimi.ai announced the release of the model weights for Kimi K3.1 Pro, along with an access link, and stated that it will continue to advance open intelligence. The model weights are now publicly available, and the official side has not provided further explanation regarding the capability adjustments.
AI-generated from public reports
Latest development · Report
Kimi Releases K3.1 Pro Model WeightsDevelopments
1 confirmed developments
Reports and discussion timeline
Browse reports and discussion by time. Their evidence levels remain distinct.
Nov 5
Kimi.aiFirst-partySelected
Kimi Releases K3.1 Pro Model Weights
Kimi announced the release of the model weights for Kimi K3.1 Pro, along with a link to obtain them, and stated that it will continue to advance open intelligence.
Nov 4
X.PINSelected
Kimi K3.1 Pro proactively weakens attack capabilities; China plans to include large models in export controls
When releasing Kimi K3.1 Pro, Moonshot AI removed cyberattack-related content from its training data and reinforcement learning tasks; in internal tests, the pass rate for exploiting software vulnerabilities dropped from 90.5% to 81.0%, in order to reduce the risk of misuse after open-weight release. On October 16, China's Ministry of Commerce released a draft update to the catalog of technologies subject to export restrictions, planning to include large models, requiring companies to register in advance, with open-source projects eligible for expedited approval. The article also mentions that when relevant UK and US institutions previously tested Kimi K3, they found that its safety rules failed to block attack tasks, but its capabilities still lagged significantly behind leading US models.