Roar AI
DiscordGet started
All models
Moonshot AI · Open weights

Kimi K3

Kimi K3 costs $3.705 per million input tokens and $18.525 per million output tokens on Roar AI. Moonshot AI’s 2.8-trillion-parameter open model, first at web research when it launched.

Input$3.705per 1M tokens
Output$18.525per 1M tokens
Context window1.05Mtokens

Kimi K3 API pricing

Input$3.705 per 1M tokens
Output$18.525 per 1M tokens
Cached input$0.351 per 1M tokens

Example: 1,000 requests with 3,000 input and 1,000 output tokens each cost $29.64. Every rate is published and stays the same from one request to the next. See every model.

Rates as of 20 November 2026.

Specs

Made byMoonshot AI
ReleasedJuly 2026
TypeChat
Context window1.05M tokens
ReadsText, images, video
Thinks before answeringAlways on
WeightsOpen (Kimi K3 License)
Size2.8T total, 104B active
Processed inSingapore
Model idkimi-k3

Is Kimi K3 worth the price?

For research agents, it is a strong candidate. Kimi K3 placed first on BrowseComp at launch with 91.2%, and it reads images and video as well as text. Its launch scores mix Moonshot’s own tests with differing agent set-ups, so try it on your own research tasks.

For everyday coding and quick answers, start elsewhere. Claude Sonnet 5.5 and DeepSeek-V4-Pro both cost less today, and Kimi K3 always thinks before it answers, so it has no fast mode for simple requests. Its open weights matter only if you plan to run it yourself, which takes a cluster of 64 or more accelerators.

Kimi K3 and the models to compare it with, priced per 1M tokens
ModelInputOutputChoose it for
Kimi K3$3.705$18.525Web research agents.
Claude Sonnet 5.5$2$10Everyday coding, documents and faster answers.
DeepSeek-V4-Pro$1.56$3.12Coding and maths, under the MIT licence.

Prices per 1M tokens.

About Kimi K3

Kimi K3 was billed as the first open model in the three-trillion class: 2.8 trillion parameters, with 104 billion working on each token. Its strongest result is web research, with 91.2% on BrowseComp, first at launch. It placed third on the Artificial Analysis Intelligence Index, behind only Claude Fable 5 and GPT-5.6 Sol.

It reads images and video as well as text, holds a million tokens, and always thinks before answering. Running it yourself takes a cluster of 64 or more accelerators, which is why most people use it through an API.

What is Kimi K3 good for?

  • Research agents that browse and read many sources.
  • Long coding agents.
  • Knowledge work across long documents and video.

Kimi K3 limitations

  • Always reasons, so there is no fast mode for simple requests.
  • Its launch scores mix Moonshot’s own tests with differing agent set-ups.
  • Moonshot’s licence asks very large products built on it to display “Kimi K3”.

How to use the Kimi K3 API

Use the OpenAI SDK you already have: point it at Roar AI and set the model name. Read the docs.

from openai import OpenAI

client = OpenAI(base_url="https://api.roar-ai.com/v1", api_key="roar_live_…")

reply = client.chat.completions.create(
    model="kimi-k3",
    messages=[{"role": "user", "content": "Summarise this contract in plain English."}],
)

Kimi K3 FAQ

How much does Kimi K3 cost?

Input is $3.705 per 1M tokens; output is $18.525 per 1M tokens. For example, 1,000 requests with 3,000 input and 1,000 output tokens each cost $29.64.

Can I use Kimi K3 alongside other models?

Yes. Every model on Roar AI is on the same API key and the same monthly invoice, so switching from Kimi K3 to another model only means changing the model name in your code.

Is Kimi K3 open source?

Its weights are open under the Kimi K3 License licence, so you could run it yourself. Through Roar AI you call it like any other model, with no servers to manage.

What is the context window of Kimi K3?

Kimi K3 can read up to 1.05M tokens in one request.

Can Kimi K3 read images and PDFs?

It reads images directly. You can still send it a PDF: Roar AI turns each page into text first, at no extra charge.

Start using Kimi K3

Create a free account, get an API key and call Kimi K3 with the OpenAI SDK.