Model overview
Headline rates land exactly on the coding tier's: $0.95 per million input tokens, $4 output, the same 262,144-token window, so no arithmetic settles this choice. Cache hits are the single divergence, $0.16 against $0.19, three cents cheaper per million, which only registers at volume on heavily reused prefixes. The separation is scope. K2.6 is offered as general-purpose multimodal work covering instruction following, extraction, dialogue and agent tasks, where the newer sibling is framed narrowly around programming. Reasoning, tool calls and context caching come either way. For mixed traffic with occasional code generation, the generalist costs the same as the specialist and targets the blend you actually send.
Billing conditions: Standard Kimi API price. Cache-miss input is shown as Input; automatic cache-hit input is shown as Cached input.
- Context window
- 262,144 tokens
- Knowledge cutoff
- —
- Price last verified
- First observed
Descriptions, capabilities, and tariffs come from the linked source. First-party prices take priority; OpenRouter fallbacks are labelled.