Skip to content

Run Kimi K3 With Claude Code on Windows or WSL

Connect Claude Code to Kimi K3 using Moonshot's Anthropic-compatible endpoint, current model ID, reasoning controls, and verification steps.

MGMCSA Guru Team August 29, 2026 3 min read
Claude Code connected to Kimi K3 through Moonshot AI

Kimi K3 is Moonshot AI’s current flagship model. It supports native vision, a 1-million-token context window, and long coding or reasoning sessions. Moonshot also provides an Anthropic-compatible endpoint, so Claude Code can connect without a translation proxy.

Install Claude Code and create a key

Follow the current Claude Code Windows and WSL setup, then create a key in the Moonshot API platform.

Before changing providers, close existing Claude Code sessions. An older provider variable in the launching terminal can override the settings you add next.

Configure Kimi K3

Open ~/.claude/settings.json in WSL or your user-level Claude settings file on Windows. Merge this env object with the settings already there:

{
  "env": {
    "ANTHROPIC_BASE_URL": "https://api.moonshot.ai/anthropic",
    "ANTHROPIC_AUTH_TOKEN": "YOUR_MOONSHOT_API_KEY",
    "ANTHROPIC_MODEL": "kimi-k3",
    "ANTHROPIC_DEFAULT_OPUS_MODEL": "kimi-k3",
    "ANTHROPIC_DEFAULT_SONNET_MODEL": "kimi-k3",
    "ANTHROPIC_DEFAULT_HAIKU_MODEL": "kimi-k3"
  }
}

Verify the model

Open a new terminal, enter a project, and start Claude Code:

cd /path/to/project
claude

Inside Claude Code, run:

/status
/model

The status should display Moonshot’s Anthropic endpoint, and the active model should be kimi-k3. Test one small repository task before allowing a broad edit.

Reasoning and multimodal input

Moonshot exposes low, high, and max reasoning effort for K3. Use lower effort for routine requests and higher effort when a task needs deeper investigation. Claude Code or the provider integration may decide which controls are available in the interface.

K3 is natively multimodal, but client support still matters. If Claude Code or an extension strips an attached image before sending the request, the model cannot inspect it. Test a simple screenshot and verify that the response refers to visible details.

Troubleshooting

An authentication error usually means the key is missing, malformed, or overridden by a shell variable. A model error usually means the old K2 ID is still present somewhere. Search the settings file and current environment for older kimi-k2 values.

If /status shows Anthropic rather than Moonshot, close Claude Code, open a clean terminal, and start it again. If a multi-turn session behaves strangely after changing models, begin a new session instead of switching to K3 midway through an existing conversation. Moonshot recommends preserving the complete assistant message in multi-turn API histories.

Kimi K3 Claude Code check

  • Create a Moonshot API key
  • Use https://api.moonshot.ai/anthropic
  • Set the model ID to kimi-k3
  • Open a new terminal and confirm /status and /model
  • Test image input separately before relying on it

Kimi K3 replaces the old K2 Thinking setup as the current flagship path. Keep the exact model ID and endpoint together, and start a fresh session when moving an existing Claude Code configuration to K3.

For other clients, see Kimi K3 with OpenCode and Kimi with Codex CLI.

Frequently asked questions

What is the current Kimi model ID?

Use kimi-k3 for Moonshot's current flagship model.

Can Claude Code connect to Kimi without a router?

Yes. Moonshot provides an Anthropic-compatible endpoint at https://api.moonshot.ai/anthropic.

Does Kimi K3 support images?

Yes. Moonshot describes K3 as natively multimodal. The client must preserve structured image content for the model to receive it.

How do I control Kimi K3 reasoning?

K3 supports low, high, and max reasoning effort. The available control depends on what the client exposes.

Sources & further reading

Official vendor documentation referenced while writing this guide.

MG

MCSA Guru Team

IT & Systems Administration

We are working IT pros and system administrators who spend our days in Windows Server, Microsoft 365, and the wider Microsoft stack. MCSA Guru is where we write down the fixes and walkthroughs we wish we had found the first time.

MCSA Guru provides independent, educational IT guidance. Microsoft, Windows, Windows Server, Microsoft 365, Exchange, and Microsoft Teams are trademarks of Microsoft Corporation; Docker is a trademark of Docker, Inc. MCSA Guru is not affiliated with or endorsed by Microsoft or Docker. Always test changes in a safe environment before applying them in production.

Related guides

Fixing something right now?

Jump straight into the guide library or search for the exact error or task you are dealing with.