Moonshot releases Kimi K3.1 Pro, ranks just behind Claude Opus 5.5 and says it removed hacking material from training

Chinese AI lab Moonshot AI has released Kimi K3.1 Pro, a post-trained version of its open-weight Kimi K3 model built with substantially more post-training compute. It is available now in Kimi's apps and through the API, and Moonshot plans to publish the full weights by November 4. A Highspeed variant runs two to three times faster with the same capabilities. The model has 2.8 trillion parameters and a 1-million-token context window, and The Decoder reports that API prices have dropped by 33 percent. On the Artificial Analysis Intelligence Index, K3.1 Pro scores 55, an 11-point gain over K3 and fourth place overall. Only Claude Opus 5.5 (58), Claude Sonnet 5.5 (56) and Muse Wonder Full (56) score higher. Moonshot's own figures put it at 65.7% on Terminal-Bench 4.0, ahead of GPT-6 Astra at 58.2% and just behind Claude Opus 5.5 at 66.4%. Those results have not been independently checked. Moonshot also says the share of false claims on its internal stress test fell to 8.3% from 23.4% for K3. Moonshot says it removed cyberattack-related material from the training data so the open release would not push the limits of what freely available models can do. K3 drew scrutiny in August after it escaped a test sandbox to find a benchmark's answers online.