Model by Moonshot AI

Kimi K2.6

Kimi K2.6 is Moonshot AI's open-weight multimodal model for long coding runs and agent setups, and it is on its way out of Moonshot's own API.

Checked 2026-11-05

Kimi K2.6 is Moonshot AI's open-weight multimodal model for long coding runs and agent setups, and it is on its way out of Moonshot's own API. The Hugging Face card gives a mixture-of-experts design with 1 trillion parameters, 32 billion active per token, 384 experts, native INT4 quantisation and a 256K context, which OpenRouter and Moonshot list as 262,144 tokens. The licence is Modified MIT; we did not read its conditions, so check them before commercial use. Moonshot's docs say K2 series models close to new sign-ups from 31 October 2026 and are to be retired by the end of December 2026, pointing users to Kimi K3.1 Pro and K3. We could not read Moonshot's price table. OpenRouter lists $0.65 per million input tokens and $3.41 per million output tokens, text and image input, from 20 April 2026, on its cheapest route as listed on 5 November 2026. Figures come from the Hugging Face card, Moonshot's docs and the OpenRouter model list. We have not run it ourselves. We would only self-host it, and use K3.1 Pro through the API.

Mentioned here

  • Kimi K3.1 Pro

    Kimi K3.1 Pro is Moonshot AI's flagship K3.1 model, a post-trained successor to Kimi K3.

    • Closed

    Moonshot AI, 1M context, $2 in / 1M

  • Kimi K3

    Kimi K3 is an open-weights model from Moonshot AI, released 2026-07-16, with a 1,048,576-token context and text, image and video input.

    • Open weights

    Moonshot AI, 1M context, $0.19 in / 1M

  • Kimi K2.7 Code

    Kimi K2.7 Code is Moonshot AI's coding-focused model, aimed at people who want an open-weight agent model that also reads images.

    • Open weights
    • Coding

    Moonshot AI, 262K context, $0.60 in / 1M

  • Moonshot AI

    Moonshot AI sells the Kimi API through its developer platform, now at platform.kimi.ai, and it suits teams who want Kimi K3 and K3.1 Pro with a 1,048,576-token context window straight from the maker.

    • Model / inference API
    • OpenAI compatible

  • OpenRouter

    OpenRouter routes more than 600 models through one API key.

    • Model gateway (router)
    • OpenAI compatible

  • vLLM

    vLLM is a high-throughput, memory-efficient serving engine for large language models, run as a server you host yourself.

    • Self-hosted runtime
    • OpenAI compatible
    • Self-hostable

Questions

What is the context window of Kimi K2.6?

262,144 tokens, with up to 235,929 tokens of output.

How much does Kimi K2.6 cost?

$0.65 per million input tokens and $3.41 per million output tokens, checked 31 October 2026.

Can I run Kimi K2.6 on my own hardware?

Yes, the weights are published under the Open-weights model licence. They are on huggingface.co.

Who makes Kimi K2.6?

Moonshot AI makes it, and released it on 20 April 2026.

Reviews

Loading reviews…

Write a review

Your rating

Up to 1000 characters. Stars alone are enough.