Moonshot AI Releases Kimi K3.1 Pro with 1M Context Window
Moonshot AI has introduced Kimi K3.1 Pro and Highspeed, featuring a 1-million-token context window and thinking-mode architecture. Available now via API and cli
Moonshot AI has officially introduced Kimi K3.1 Pro on October 24, 2026, marking the debut of its new K3.1 model family. Featuring a massive 1-million-token context window, the release is accompanied by a faster companion variant named Kimi K3.1 Pro Highspeed, designed to deliver high-throughput reasoning capabilities.
Image source: https://www.unite.ai/moonshot-ai-releases-kimi-k3-1-pro-first-model-in-new-k3-1-series/
Kimi K3.1 Pro operates exclusively in a thinking mode optimized for deep reasoning and complex execution workflows. According to Moonshot AI, the model scored 65.7% on Terminal-Bench 4.0—outperforming GPT-6 Astra at 58.2% and approaching Claude Opus 5.5 at 66.4%. In addition, it achieved 92.7% with tool use on BenchCAD, a benchmark evaluating CAD code generation and 3D reconstruction. The Highspeed variant matches these core capabilities while delivering two to three times the execution speed of the standard Pro model.
Both models are immediately accessible across the Kimi ecosystem, including Kimi.ai, mobile apps for iOS, Android, and HarmonyOS, the Kimi Work desktop application (version 3.1.0 or later), Kimi Code, and the Kimi API. Enterprise API users also benefit from a Zero Data Retention option. However, the complete model weights are scheduled for public release on November 4, 2026, meaning open-weights deployment follows the initial platform rollout.
This launch underscores the rapid expansion of ultra-long-context reasoning models in production environments. Users should note that K3.1 Pro requires thinking mode enabled for requests, and accessing the full 1M context or Highspeed tier requires corresponding subscription levels.