Seldon Launches VGI-Bench for Multimodal AI

Seldon released VGI-Bench, a new multimodal benchmark with 550 questions to evaluate 12 visual and audio-visual capabilities. Results show top AI models barely pass, while humans achieve 84.5% accuracy, highlighting a significant gap between AI and human performance.

2026-08-08 ~ 2026-08-08 · 2 related posts

• VGI-Bench Multimodal Eval: Best Model Only 64.73% vs Humans at 84.5% — lateinteraction · 2026-08-08

1 near-duplicate retellings: bclavie