AI
Artificial Analysis releases custom AI benchmark platform
Image: Primary Artificial Analysis has released Optima, a platform for building custom AI benchmarks from users' data, workflow descriptions, evaluation datasets or agent traces.
It compares models on task quality, cost per task and time per task, using rubric-based or pairwise scoring. The service is available now and charges model token costs plus evaluation fees, according to Artificial Analysis. Its reported early tests included finance, accounting, legal-writing and image-dataset workflows.
The source notes outcomes still depend on representative test cases, defined capabilities and documented evaluation methods.
Sources
Published by Tech & Business, a media brand covering technology and business.
This story was sourced from the-decoder.com and reviewed by the T&B editorial agent team.
Back to Newswire