Kimi K3.1 Pro
Moonshot AI’s open-weight multimodal model for coding, financial research, 3D creation, and long-horizon agentic work, with a one-million-token context window.
Model
Kimi K3.1 Pro builds on Kimi K3 with substantially more post-training compute and large-scale reinforcement learning across hundreds of domains.
For software engineering, it handles codebase fixes and full-project builds, including web interfaces, games, and interactive 3D scenes. Its agentic training strengthens multi-step execution, tool use, and iterative testing and debugging.
For financial work, it retrieves and reconciles evidence across documents, performs multi-step calculations, and builds financial models and spreadsheets. Its 3D capabilities include generating Blender and CAD scripts from text briefs, sketches, and reference images.
Moonshot reports 65.7% on Terminal-Bench 4.0, 92.7% on BenchCAD with tools, and 64.4 on OfficeQA Pro v2. Full results and evaluation settings are available in Moonshot’s announcement.
Shared Endpoint
On a Shared Endpoint, you pay per token. The endpoint is OpenAI-compatible and already live: point your existing SDK at it and start sending requests.
modal endpoint create --shared --model moonshotai/Kimi-K3.1-Pro