Codex Configuration

June 29, 2026 · View on GitHub

Codex Configuration

Pay-as-you-go MiMo API and Token Plan both support Codex. Refer to this guide for configuration and usage.

MiMo now supports the Responses API. Users who are using the Chat Completions API can follow this document to update.

Prerequisites

Obtain Credentials

Two usage methods are supported, but the corresponding credential acquisition methods differ:

Usage Method	Description	Acquisition Method
Pay-as-you-go API	Charged based on actual usage, suitable for light use	BASE_URL: `https://api.xiaomimimo.com/v1` / API Key format: `sk-xxxxx` — Go to API Keys to create an API Key
Token Plan	Fixed subscription fee, with limited calls based on the package	BASE_URL: `https://token-plan-cn.xiaomimimo.com/v1` / API Key format: `tp-xxxxx` — After successful subscription, go to Token Plan to obtain the exclusive Base URL and API Key

Install Codex

Prerequisites: Node.js 18 or a later version must be installed first.

Installation command:

npm install -g @openai/codex

Verify the installation (a version number output indicates success):

codex --version

Edit Configuration File

Note

When configuring basic information, first check if the MIMO_API_KEY environment variable exists. If it does, please clear it or replace the value with the API Key obtained through the corresponding usage method.

To switch models directly via the /model command, refer to the "Further Reading - Custom MiMo Model Configuration" section and configure the model-catalogs.json file.

Configuration file paths:

macOS/Linux: ~/.codex/config.toml
Windows: User directory\.codex\config.toml

Below are the complete configuration examples. The model field can be changed to other supported models (e.g., mimo-v2.5) as needed.

Pay-as-you-go

1. Edit or create the configuration file config.toml

model = "mimo-v2.5-pro"
model_provider = "mimo"
model_reasoning_effort = "high"

# Enable model reasoning summaries; if set to false, model_reasoning_effort will not take effect even if configured
model_supports_reasoning_summaries = true
model_reasoning_summary = "none"
model_context_window = 1048576

web_search = "disabled"

[model_providers.mimo]
name = "mimo"
base_url = "https://api.xiaomimimo.com/v1"
env_key = "MIMO_API_KEY"
wire_api = "responses"

2. Configure the environment variable MIMO_API_KEY

Pay-as-you-go uses an API Key starting with sk-.

macOS/Linux

echo 'export MIMO_API_KEY="sk-your-api-key-here"' >> ~/.bashrc
source ~/.bashrc

Windows (CMD)

# Run the following command in CMD
setx MIMO_API_KEY "sk-your-api-key-here"

# After success, open a new command prompt and run the following to verify the variable is set.
echo %MIMO_API_KEY%

Token Plan

1. Edit or create the configuration file config.toml

model = "mimo-v2.5-pro"
model_provider = "mimo"
model_reasoning_effort = "high"

# Enable model reasoning summaries; if set to false, model_reasoning_effort will not take effect even if configured
model_supports_reasoning_summaries = true
model_reasoning_summary = "none"
model_context_window = 1048576

web_search = "disabled"

[model_providers.mimo]
name = "mimo"
base_url = "https://token-plan-cn.xiaomimimo.com/v1"
env_key = "MIMO_API_KEY"
wire_api = "responses"

2. Configure the environment variable MIMO_API_KEY

Token Plan uses an API Key starting with tp-.

macOS/Linux

echo 'export MIMO_API_KEY="tp-your-api-key-here"' >> ~/.bashrc
source ~/.bashrc

Windows (CMD)

# Run the following command in CMD
setx MIMO_API_KEY "tp-your-api-key-here"

# After success, open a new command prompt and run the following to verify the variable is set.
echo %MIMO_API_KEY%

Use Codex CLI

After completing the above configuration, open a new terminal and run the following command to start Codex.

codex

Codex provides a VS Code plugin. Search for "Codex" in the VS Code Extension Marketplace to install it. The plugin automatically reuses the existing local Codex configuration. If you have never used the Codex command-line tool, please follow the specifications in the "Edit Configuration File" section to complete the configuration.

Field	Description
`slug`	Unique internal identifier for the model; must exactly match the backend API model name
`display_name`	Model name displayed in the frontend UI; can be the same as slug or a custom display name
`description`	Model description text, used for hover tooltips in the model list and model detail views
`default_reasoning_level`	Default reasoning intensity level for new sessions. For MiMo models, `none` disables thinking; other values enable thinking
`supported_reasoning_levels`	List of reasoning intensity configurations available for users to switch between, containing effort identifiers and description text. `none` and `high` are provided as examples
`shell_type`	Tool execution environment type. The example uses `"shell_command"`, indicating support for Shell command tools
`visibility`	Model visibility strategy in the client UI; `list` means normal display
`supported_in_api`	Whether the model is open for standard API calls; `true` means external programs can call it via API
`priority`	UI sorting weight; smaller numbers appear earlier, `1` represents high priority pinned at the top
`base_instructions`	System prompt
`supports_reasoning_summaries`	Whether the model supports reasoning summary output. When `true`, requests carry reasoning parameters; when `false`, all reasoning fields are cleared and the reasoning level (`reasoning_level`) configuration becomes ineffective
`default_reasoning_summary`	Default reasoning summary output mode; `none` means no separate reasoning summary is output
`support_verbosity`	Whether output verbosity control is supported; `false` means text verbosity level switching is not yet supported
`truncation_policy`	Context truncation strategy; mode is the truncation unit, limit is the maximum byte limit per context
`supports_parallel_tool_calls`	Whether parallel multi-tool calls are supported; `false` means only sequential calls are supported
`supports_image_detail_original`	Whether original high-resolution image parsing is supported for image input; `true` means using the full original resolution
`context_window`	Model nominal total context window size; `1048576` = 1M context window
`max_context_window`	Maximum supported context window limit
`effective_context_window_percent`	Actual usable context percentage; `95` means reserving a 5% safety buffer to prevent exceeding the limit
`experimental_supported_tools`	Experimental feature tool whitelist; an empty array `[]` means no additional beta tools are currently available
`input_modalities`	Supported input modality types; `["text", "image"]` means text + image input is supported
`supports_search_tool`	Whether built-in web search tool capability is supported; `false` means web search is not supported

Configuration Steps

Edit or create the ~/.codex/model-catalogs/model-catalogs.json file for detailed MiMo model configuration (Take mimo-v2.5-pro and mimo-v2.5 as examples)

{
  "models": [
    {
      "slug": "mimo-v2.5-pro",
      "display_name": "mimo-v2.5-pro",
      "description": "MiMo-v2.5-Pro: Trillion-parameter Flagship Agent Foundation",
      "default_reasoning_level": "high",
      "supported_reasoning_levels": [
        {
          "effort": "none",
          "description": "Disable Thinking"
        },
        {
          "effort": "high",
          "description": "Enabled Thinking"
        }
      ],
      "shell_type": "shell_command",
      "visibility": "list",
      "supported_in_api": true,
      "priority": 0,
      "base_instructions": "You are MiMo, an AI assistant developed by Xiaomi. Today's date: {date} {week}. Your knowledge cutoff date is December 2024.",
      "supports_reasoning_summaries": true,
      "default_reasoning_summary": "none",
      "support_verbosity": false,
      "truncation_policy": {
        "mode": "bytes",
        "limit": 10000
      },
      "supports_parallel_tool_calls": false,
      "supports_image_detail_original": false,
      "context_window": 1048576,
      "max_context_window": 1048576,
      "effective_context_window_percent": 95,
      "experimental_supported_tools": [],
      "input_modalities": ["text"],
      "supports_search_tool": false
    },
    {
      "slug": "mimo-v2.5",
      "display_name": "mimo-v2.5",
      "description": "MiMo-V2.5: Native Omni-modal Perception Model",
      "default_reasoning_level": "high",
      "supported_reasoning_levels": [
        {
          "effort": "none",
          "description": "Disable Thinking"
        },
        {
          "effort": "high",
          "description": "Enabled Thinking"
        }
      ],
      "shell_type": "shell_command",
      "visibility": "list",
      "supported_in_api": true,
      "priority": 1,
      "base_instructions": "You are MiMo, an AI assistant developed by Xiaomi. Today's date: {date} {week}. Your knowledge cutoff date is December 2024.",
      "supports_reasoning_summaries": true,
      "default_reasoning_summary": "none",
      "support_verbosity": false,
      "truncation_policy": {
        "mode": "bytes",
        "limit": 10000
      },
      "supports_parallel_tool_calls": false,
      "supports_image_detail_original": true,
      "context_window": 1048576,
      "max_context_window": 1048576,
      "effective_context_window_percent": 95,
      "experimental_supported_tools": [],
      "input_modalities": ["text", "image"],
      "supports_search_tool": false
    }
  ]
}

Add the following configuration at the root level of the config.toml file

macOS/Linux

model_catalog_json = "~/.codex/model-catalogs/model-catalogs.json"

Windows

model_catalog_json = "User directory/.codex/model-catalogs/model-catalogs.json"

Resources

MiMo Codex Configuration Guide
Codex CLI — OpenAI's coding agent.
MiMo Official Website
MiMo Platform — API key management and usage.

Codex Configuration

Codex Configuration

Prerequisites

Obtain Credentials

Install Codex

Edit Configuration File

Pay-as-you-go

Token Plan

Use Codex CLI

Use Codex IDE Plugin

Further Reading

Custom MiMo Model Configuration

Key Field Descriptions

Configuration Steps

Resources