Scale AI Launches SEAL Showdown Benchmarking System

Scale AI launched SEAL Showdown, a public benchmarking system that ranks artificial intelligence models by usability across various professions, age groups, and regions. The system utilizes voluntary votes from contributors in more than 100 countries, including physicists, lawyers, and physicians, to offer a more representative alternative to standard global averages.

The company positions the tool as a resource for businesses to select models tailored to specific tasks and markets, placing it in direct competition with platforms like LMArena. Early data from the system shows that Claude leads in reasoning and writing, while ChatGPT is more effective for classification and brainstorming. Google's Gemini performs better for users over the age of 50.

Product manager Janie Gu noted that participation in the system is voluntary and unpaid, stating that contributors view the opportunity as a perk because they receive free access to frontier models that typically cost hundreds of dollars a month.