MoonshotAI: Kimi K2 0711

Kimi K2 0711 is a text-only model from MoonshotAI with a 131K token context window and a 32K token output cap. It supports tool use, which makes it usable for agentic and function-calling workflows, but it does not offer native reasoning mode and structured output support is unconfirmed. Input is limited to text, so multimodal tasks are out of scope. At $0.57 per million input tokens and $2.30 per million output tokens, it sits in a mid-range price band. Its blended benchmark score of 65.8 comes from a single benchmark (aider_polyglot at 59.1), so that figure reflects a narrow slice of performance rather than a broad evaluation. Buyers running code-generation or agentic pipelines who want tool support at a moderate price may find it worth testing, but the thin benchmark coverage means direct comparisons to better-tested models carry real uncertainty.