Playground
Test your prompts
Run one prompt on up to three models side by side. Compare the answers, speed and cost before you pick one.
What you see
Compare models with one prompt
Send the same prompt to up to three models at once and read their replies side by side.
Speed and cost per reply
Time to first token, tokens per second and the charge for each run.
Real requests
Each run is a real API request on one of your keys.
Nothing stored
Prompts and replies are gone when you leave the page.
When to use it
Picking a model for a new agent
Run a real prompt on two or three candidates before you commit.
Checking a cheaper model
Put it next to the one you pay for. Compare the answers and the charge.
Trying a newly added model
Run your own prompt on it before you change your code.
Playground questions
Is the playground free?
No. Each run is a real request at the model's list price.
Can I see the cost before I send?
Yes. The playground quotes the cost before the run goes out.
Are my prompts saved?
No. Hopscotch never stores prompts or responses by default. Leaving the page clears them.
Do I need an API key?
Yes. Each run uses one of your workspace keys, so its spend shows on that key. A workspace with no keys creates one first.
Run your prompt on three models
Open the playground and paste a prompt from production.