MTS launches Creative Benchmarks to test AI on open-ended creative tasks
Overview
MTS introduces Creative Benchmarks, a set of tests that evaluate AI models on open-ended creative work.
Example tasks include teaching history as an anime, building a game players want to keep playing, and making people feel something. The company says the benchmark targets creative tasks with no single right answer, unlike most existing AI evaluations. Details are published at drops.mts.now/creative.
Written by AI from the articles below · updated Oct 13, 6:41 PM ET
Check the sources:
Article timeline
The articles in this story. Times are ET.
MTS@MTSliveXMTS launches Creative Benchmarks testing AI on open-ended creative tasksAIMTS introduces Creative Benchmarks, which test AI models on open-ended tasks such as teaching history as an anime, building a game players want to keep playing, and making people feel something. The benchmark targets creative work with no single right answer, unlike most existing evaluations. Details are available at drops.mts.now/creative.

Heat trend
Not enough continuous observations to show a trend yet.