Tools
Free in your browser. Nothing to install. Made by a Rudder.chat user
Choose from skills such as reasoning, logic, coding, math, science, writing and IT skills, each with sample questions. Each answer gets scores for accuracy, reasoning and clarity, plus strengths and issues.
Enter one question, choose a model for each of two pasted answers, and see which one is stronger with a score for each. Past comparisons are saved and can be exported as a CSV file.
Try an AI agent on tasks like refund requests, meeting scheduling, IT troubleshooting and security reviews. You can also review your past runs and category averages, and see a shared leaderboard of scores.
Step 1
Choose a skill, then pick a sample question or type your own.
Step 2
Run the benchmark and read the answer, the score and the evaluation notes.
Step 3
Open the compare tab to judge two answers, or check your history and the leaderboard.
Most tests are scored on accuracy, reasoning and clarity, combined into an overall score out of 100. IT skills questions use diagnosis, safe troubleshooting, practical steps and escalation instead.
Yes. Paste the same question and one answer from each model, choose the model names from the list, and the judge scores both and picks a winner or a tie.
No. The shared leaderboard shows only scores and the skill category, not your questions or answers. Your history and saved comparisons are visible only to you.
Make your own mini app
Describe it in the chat
