arize-experiment

Run and compare model evaluations in Arize.

model evaluation

What it does

Runs, manages, and analyzes Arize experiments for evaluating model performance.

When to use it

Use when the user wants to run experiments, compare models, evaluate AI, benchmark results, or measure accuracy.

Give it an Arize experiment task; it creates, runs, manages, or analyzes experiments in the user's account.

This skill

Arize experiments

Write

Changes and deletes Arize experiments (irreversible)

target model API

Write

Sends examples to model APIs

Arize experiment data

Read

Reads your Arize experiment data

target provider's SDK

Execute

Installs the model provider SDK

Requires ax CLI version 0.33.0 or later.

Requires access through a configured Arize profile.