Need an AI that can actually read your whole document?

A long-context AI assistant from Moonshot that reads whole codebases, researches on its own, and runs agentic tasks for you

Quick answer: I keep Kimi around as my second chatbot for a reason.

Our Verdict

I keep Kimi around as my second chatbot for a reason. When a document is too long for a free-tier chat window or I want an agent to chew through a research job on its own, Kimi's long context and credit-priced plans beat paying $20 a month for occasional heavy lifting. Just keep sensitive code off the free plan, and expect the occasional rough edge outside China.

Visit Kimi →

Pricing shown below is approximate and may have changed — always confirm on the official site.

Made this tool? Get the free "Featured on MustUseAI" badge for your site.

🚀 Key Features

Long-context chats, up to 1 million tokens on higher plans, so a whole codebase or a stack of long documents fits in one session
Deep Research that reads the web and compiles reports, plus agent multi-tasking and scheduled tasks that run while you are away
Kimi Code, Moonshot's agentic coding companion, with Goal mode on paid plans for long autonomous engineering runs
Website deployment built in: build and publish small sites with database support straight from a chat
An open-weight model family (K3 released under Moonshot's Kimi K3 License) plus a separate pay-per-token developer API

💰 Pricing

Plan	Details
Plan	Free (~$0/mo): the full assistant on the web and mobile apps with limited usage of the flagship K3.1 Pro model
Plan	Plus (~$15/mo billed annually, ~$19 monthly): more agent credits, Kimi Code coding access, Deep Research, website deployment and scheduled tasks
Plan	Pro (~$31/mo billed annually, ~$39 monthly): 2x agent credits, faster K3.1 Pro Highspeed, and Goal mode for long autonomous runs
Max (~$79/mo billed annually, ~$99 monthly): 5x agent credits plus up to 1M tokens of context for ultra	long chats
Plan	Ultra (~$159/mo billed annually, ~$199 monthly): 10x agent credits and the highest usage allowances

free tier + paid plans from ~$15/mo (approximate) A free tier is available if you want to try before you pay.

👍👎 Pros & Cons

✅ What we like

The free tier is genuinely usable, with limited access to the flagship K3.1 Pro model and no credit card required
Long-context handling is among the best in its class, and the agentic features feel native rather than bolted on
Strong value for coding work: Kimi Code, Goal mode and Swarm multi-agent runs at prices that undercut the big US labs
Open weights for self-hosting, and a cheap per-token API for developers

🤔 What could be better

Paid plans are metered in credits rather than simple message limits, which makes the real monthly cost harder to predict
The best features, 1M-token chats, Highspeed and full agent allowances, sit behind the $31/mo and higher tiers
The app and docs still lean Chinese-first in places, and some new features land there before the international rollout
Moonshot is a Chinese company with data flowing through Chinese infrastructure, so keep sensitive work code off the free plan

❓ Frequently Asked Questions

Is Kimi really free?

Yes. The Free plan costs nothing and gives you limited usage of the flagship K3.1 Pro model on the web and in the mobile apps. Paid plans buy more agent credits, faster Highspeed access, and features like 1M-token chats.

Who makes Kimi?

Kimi is built by Moonshot AI, a Beijing-based AI lab founded in 2023. The company also releases open-weight versions of its K-series models for developers and researchers.

What is Kimi Code?

Kimi Code is Moonshot's agentic coding companion: it plans, writes, tests and debugs code across multi-step runs instead of just answering questions. It is included from the Plus plan upward.

Is Kimi better than ChatGPT?

For long documents and heavy agent work, Kimi's context window and pricing are hard to beat. For everyday chat, polish and third-party integrations, ChatGPT still has the edge. Many people use both side by side.