You can deploy this model and start earning money today!

Views

No views yet

a benchmark for evaluating AI agents as meeting participants by Carcinize Corp · April 2026 · companion to Modyl (meeting-agent product) overview · dataset · leaderboard

quick start

Three commands. Install, evaluate, report. pip install modylbench

score pre-recorded agent responses (see response format below)

modylbench evaluate --responses responses.jsonl --output scorecard.json

see all available scenariosmodylbench list-scenarios

Or load the… See the full description on the dataset page: https://huggingface.co/datasets/carcinize/modylbench.