Kimi K2.6

MoonshotAI· released 20 Apr 2026

moonshotai/kimi-k2.6
Input / 1M tokens
$0.434
Output / 1M tokens
$1.828
Context window262K tokens

Longest answer: 236K tokens

reads textreads imageopen weightstool callingvisionreasoning

About

Kimi K2.6 is a multimodal model from MoonshotAI for long-running coding work, coding-led interface generation, and coordinating multiple agents. It is described as handling complex, end-to-end coding tasks in Python, Rust, and Go. The model accepts text and images and returns text, with vision and reasoning support. Its context window is 262,144 tokens, and its maximum output is 235,929 tokens. Kimi K2.6 has open weights; its Hugging Face identifier is moonshotai/Kimi-K2.6. The model is available as moonshotai/kimi-k2.6. Pricing is $0.4341 per 1M input tokens and $1.828 per 1M output tokens. It was released on April 20, 2026.

Who it is for

Kimi K2.6 suits teams building coding workflows that span Python, Rust, or Go, particularly where tasks involve interface generation or multiple agents. Its image input and long context may fit projects that need multimodal prompts and substantial context.

What is good

  • Open weights are available.
  • Accepts both text and image inputs.
  • Context window is 262,144 tokens.
  • Supports reasoning and tools.
  • Maximum output is 235,929 tokens.

What to know first

  • Input costs $0.4341 per 1M tokens.
  • Output costs $1.828 per 1M tokens.

Verdict

Kimi K2.6 combines long-context multimodal input with a focus on extended coding tasks and agent coordination. Check the listed token prices and open-weight availability against your deployment requirements.

Details

Lab
MoonshotAIopenrouter.ai · 3 Oct 2026
Context
262Kopenrouter.ai · 3 Oct 2026
Input price
$0.434 / 1Mopenrouter.ai · 3 Oct 2026
Output price
$1.828 / 1Mopenrouter.ai · 3 Oct 2026
Max output
235,929 tokensopenrouter.ai · 3 Oct 2026
Inputs
text, imageopenrouter.ai · 3 Oct 2026
Open weights
Yesopenrouter.ai · 3 Oct 2026

More longest-context models

See the list