Connect Codex CLI
Configure config.toml to switch between platform models.
Codex CLI connects through the OpenAI-compatible protocol. Add a custom model provider, then use -m to select any platform model.
Install Codex CLI
npm install -g @openai/codexAdd the GoodPut provider
model_provider = "goodput"
model = "z-ai/glm-5.2"
[model_providers.goodput]
name = "GoodPut AI"
base_url = "https://api.goodputai.cn/v1"
env_key = "GOODPUT_API_KEY"Set the key and start
export GOODPUT_API_KEY="YOUR_GOODPUT_API_KEY"
codex # Use the default model
codex -m deepseek/deepseek-v4-pro # Switch models for this runenv_key names an environment variable. Codex reads the key from that variable at startup; do not put a plaintext key in config.toml.
Use the Responses protocol (optional)
The platform also supports OpenAI Responses (/v1/responses). To use it in Codex, add this line to the provider configuration:
[model_providers.goodput]
name = "GoodPut AI"
base_url = "https://api.goodputai.cn/v1"
env_key = "GOODPUT_API_KEY"
wire_api = "responses"If a model has compatibility problems, remove wire_api to return to the default Chat Completions protocol.