Promptfoo runs prompt + model evals from a single YAML. Add AIgateway as a provider and you can grade your prompt against 5 models in one command.
Add AIgateway-hosted models to promptfooconfig.yaml.
providers:
- id: openai:chat:anthropic/claude-opus-4.7
config: { apiBaseUrl: "https://api.aigateway.sh/v1" }
- id: openai:chat:moonshot/kimi-k2.6
config: { apiBaseUrl: "https://api.aigateway.sh/v1" }
- id: openai:chat:openai/gpt-5.4
config: { apiBaseUrl: "https://api.aigateway.sh/v1" }