AIThe Next Webabout 2 hours ago
Visa AI Benchmark Fails 42.8% of Runs
3 min read
Visa launched its first AI shopping benchmark on Thursday, testing 22 models from 12 labs across 15 tasks. The AI agents failed 42.8% of their test runs, with the best performing model scoring 82.2%. While 72% of US consumers use AI assistants, only 23% trust them to make payments.
Level
Hype check
Tap to vote and see what everyone thinks.
