Model Lab

AI Playground

Experiment with prompts, compare responses, and analyze AI quality signals in one place.

Welcome to AI Playground.

Use models from the catalog, submit prompt tests, and run A/B comparisons with side-by-side metrics.

Fetching model catalog…

Prompt experiments

Test a single prompt variant with configurable model settings.

Ready to run prompt tests.

Run name Model System prompt User prompt Context / documents

Temperature: 0.70 Top-p: 1.00 Max output tokens

Prompt Test Response

No response yet. Run a prompt test to populate this panel.

Run name

Model A Model B

System prompt A System prompt B User prompt Shared context Judge criterion

Temperature: 0.70 Max output tokens

A/B Result

No comparison yet. Run an A/B test to see variant details.

Ready to compare model variants.

Loading last runs…

Time	Type	Name	Model	Status	Latency	Quality

Window

Loading analytics…

Total Runs 0

Prompt Runs 0

A/B Runs 0

Success Rate 0.0%

Avg Latency 0ms

Avg Quality 0%

Model breakdown