Moonshot AI launched Kimi K3.1 Pro, a post-trained version of K3 with a 1M token context. Its own numbers put it at 65.7% on Terminal-Bench 4.0, just behind Claude Opus 5.5 at 66.4% and ahead of GPT-6 Astra at 58.2%. Its claim level hallucination rate fell from 23.4% to 8.3%. Moonshot says it deliberately removed cyber training data, which costs about nine points on a security benchmark. The full weights are promised by 4 November. Every benchmark figure is Moonshot's own.