Using Kimi in OpenCode

OpenCode is a terminal-based coding agent. This guide shows how to forward OpenCode model requests to Kimi For Coding via built-in authentication, with Kimi K3 or Kimi K2.7 Code providing the model inference.

Install OpenCode

The easiest way to install OpenCode is through the installation script:

curl -fsSL https://opencode.ai/install | bash

Or install via npm:

npm install -g opencode-ai

Configure API key

Run opencode auth login and select Kimi For Coding from the provider list:

$ opencode auth login
┌  Add credentials
│
◆  Select provider
│  ● Kimi For Coding
│  ...
└

Enter your Kimi For Coding API Key:

$ opencode auth login
┌  Add credentials
│
◇  Select provider
│  Kimi For Coding
│
◇  Enter your API key
│  _
└

API Keys can be created and managed in the Kimi Code Console (up to 5 keys; each key is shown only once at creation).

Launch and use

After configuring, start OpenCode:

$ opencode

Type /models in a session:

/models

Pick a model under the Kimi For Coding list:

• Kimi K3 (k3): The most capable flagship model
• Kimi K3-256K (k3-256k): k3-256k is now available. Within 256k context, it delivers the same results. k3 (1M) consumes about twice as much quota as k3-256k. Ideal for everyday Q&A, code completion, routine feature development, and single-file or small-file edits — video input is not supported.
• Kimi K2.7 Code (kimi-for-coding): A mature, stable coding model
• Kimi For Coding HighSpeed (kimi-for-coding-highspeed): The high-speed version of K2.7 Code

Recommended model launch

k3-256k is now available. Within 256k context, it delivers the same results. k3 (1M) consumes about twice as much quota as k3-256k. Ideal for everyday Q&A, code completion, routine feature development, and single-file or small-file edits — video input is not supported. If switching from k3 (1M) to k3-256k, compress the context first before switching.

Before you start, check your membership tier and what it entitles you to — calling models or context windows beyond your plan will fail with an error:

Tier	Available models	Context window
Andante	Kimi K2.7 Code	256K
Moderato	Kimi K3 Kimi K3-256K Kimi K2.7 Code	256K each
Allegretto and above	Kimi K3 Kimi K3-256K Kimi K2.7 Code Kimi For Coding HighSpeed	Kimi K3: up to 1M Kimi K3-256K: 256K Kimi K2.7 Code: 256K Kimi For Coding HighSpeed: 256K

With K3 selected, type /variants to switch the thinking effort:

/variants

Thinking effort mapping

OpenCode offers the variants Default / low / high / max, mapped to K3's actual effort levels as follows (full mapping table in Model configuration):

OpenCode variant	K3 actual effort
Default	high (K3 default)
low	low
high	high (recommended)
max	max

Kimi K3 supports low / high / max; Kimi K2.7 Code series does not require this setting.

See Model configuration for model capabilities and setup details.

Next steps

• Model configuration — Compare the models and see the full effort mapping table
• Membership Benefits — Confirm which models are available for your plan