Claude Code with B.AI API

Table of Contents

• Getting Started
• Option 1: settings.json (Recommended)
• Option 2: Environment Variables
• Switching Models
• Supported Model IDs for Claude Code
• FAQ

---

Getting Started

Complete the following two prerequisites before proceeding with configuration.

Step 1: Install Claude Code

Choose the installation method for your operating system:

macOS / Linux / WSL:

curl -fsSL https://claude.ai/install.sh | bash

Windows PowerShell:

irm https://claude.ai/install.ps1 | iex

npm (requires Node.js 18+):

npm install -g @anthropic-ai/claude-code

Step 2: Obtain a B.AI API Key

1. Log in to the B.AI Chat Platform
2. Navigate to the API Key Management Page
3. Click to create a new API Key

Note: Keep your API Key secure — you will need it in the configuration steps below.

Once both prerequisites are complete, choose either option below to configure the B.AI API.

---

Option 1: settings.json (Recommended)

This approach configures Claude Code via its own settings file, without modifying system environment variables.

Steps

1. Locate the settings.json file in the Claude Code installation directory. If it does not exist, create a new one.
2. Add the following content:

{
  "env": {
    "ANTHROPIC_AUTH_TOKEN": "Your-B.AI-API-Key",
    "ANTHROPIC_BASE_URL": "https://api.b.ai",
    "ANTHROPIC_MODEL": "claude-sonnet-4-6",
    "API_TIMEOUT_MS": "3000000"
  }
}

Tip: You can replace the ANTHROPIC_MODEL value with any model from the Supported Model IDs for Claude Code list.

3. Save the file and restart your terminal.
4. Verify the configuration:

claude --version
claude

If Claude Code starts successfully and responds using the B.AI model, the configuration is complete.

---

Option 2: Environment Variables

This approach is suitable when you need the configuration to take effect globally.

Set Environment Variables

Add the following lines to the appropriate shell configuration file:

Bash (Linux / macOS)

Edit ~/.bashrc (or ~/.bash_profile on macOS):

# Claude Code - B.AI Provider Configuration
export ANTHROPIC_BASE_URL="https://api.b.ai"
export ANTHROPIC_AUTH_TOKEN="your-api-key-here"
export ANTHROPIC_MODEL="claude-sonnet-4-6"
# End of B.AI Provider Configuration

Apply the changes:

source ~/.bashrc

Zsh (Linux / macOS)

Edit ~/.zshrc:

# Claude Code - B.AI Provider Configuration
export ANTHROPIC_BASE_URL="https://api.b.ai"
export ANTHROPIC_AUTH_TOKEN="your-api-key-here"
export ANTHROPIC_MODEL="claude-sonnet-4-6"
# End of B.AI Provider Configuration

Apply the changes:

source ~/.zshrc

PowerShell (Windows)

Edit your $PROFILE file:

# Claude Code - B.AI Provider Configuration
$env:ANTHROPIC_BASE_URL = "https://api.b.ai"
$env:ANTHROPIC_AUTH_TOKEN = "your-api-key-here"
$env:ANTHROPIC_MODEL = "claude-sonnet-4-6"
# End of B.AI Provider Configuration

Apply the changes:

. $PROFILE

Configuration File Path Reference

OS	Shell	Configuration File
Linux	Bash	~/.bashrc
Linux	Zsh	~/.zshrc
macOS	Bash	~/.bashrc or ~/.bash_profile
macOS	Zsh	~/.zshrc
Windows	PowerShell	$PROFILE

Verify the Configuration

claude --version
claude

If Claude Code starts successfully and responds using the B.AI model, the configuration is complete.

---

Switching Models

Temporary Switch (Current Session Only)

Linux / macOS:

ANTHROPIC_MODEL=claude-opus-4-6 claude

Windows PowerShell:

$env:ANTHROPIC_MODEL="claude-opus-4-6"; claude

Permanent Switch

Update the ANTHROPIC_MODEL value in your settings.json or shell configuration file, then restart the terminal or reload the configuration.

---

Supported Model IDs for Claude Code

Use one of the model IDs below as the ANTHROPIC_MODEL value. These models have been verified with B.AI's Anthropic-compatible /v1/messages interface and tool-use flow for Claude Code. This list may differ from the full B.AI model catalog.

Model ID	Provider
claude-opus-4-8	Anthropic
claude-opus-4-7	Anthropic
claude-opus-4-6	Anthropic
claude-opus-4-5	Anthropic
claude-sonnet-4-6	Anthropic
claude-sonnet-4-5	Anthropic
claude-haiku-4-5	Anthropic
gpt-5.5	OpenAI
gpt-5.5-instant	OpenAI
gpt-5.4-pro	OpenAI
gpt-5.4	OpenAI
gpt-5.4-mini	OpenAI
gpt-5.4-nano	OpenAI
gpt-5.2	OpenAI
gemini-3.5-flash	Google
gemini-3-flash	Google
deepseek-v3.2	DeepSeek
kimi-k2.5	Moonshot
glm-5.1	Zhipu AI
minimax-m3	MiniMax
minimax-m2.7	MiniMax

Some B.AI models may be available for normal API calls but are not listed here because Claude Code requires reliable tool-use behavior. For Claude models, both hyphenated aliases and dot-version aliases are accepted in Claude Code configuration.

---

FAQ

Q: I get "claude command not found" — what should I do?

Install Claude Code first:

• Linux / macOS: curl -fsSL https://claude.ai/install.sh | bash
• Windows: irm https://claude.ai/install.ps1 | iex

Restart your terminal after installation.

Q: The setup script fails — what should I check?

Verify the following:

1. Claude Code is installed (claude --version outputs a version number)
2. curl is available (Linux / macOS)
3. Your network can reach https://api.b.ai

If the issue persists, try Option 2: Environment Variables for manual configuration.

Q: How do I verify the configuration is working?

Check that the environment variables are set:

Linux / macOS:

echo $ANTHROPIC_BASE_URL
echo $ANTHROPIC_MODEL

Windows PowerShell:

$env:ANTHROPIC_BASE_URL
$env:ANTHROPIC_MODEL

Q: How do I check which model is currently in use?

After launching Claude Code, type:

/status

Q: Why do I get 403 access_denied when calling some models?

Models with a Premium tag may return 403 access_denied if your B.AI account has not been recharged before.

A valid API key alone is not sufficient for Premium models. Please recharge your account first, then try again.

Q: How do I remove the B.AI provider configuration?

Edit your shell configuration file and delete everything between the following comment blocks:

# Claude Code - B.AI Provider Configuration
...
# End of B.AI Provider Configuration

Then reload the configuration file (e.g., source ~/.zshrc).

---

Version: v1.0 | Last Updated: 2026-03-27