FrontierRSI Benchmark: Top AI Models Boost Research Productivity by Only ~34%

Abundant Labs, founded by ex-Waymo engineers, released the FrontierRSI benchmark testing AI on real frontier lab research tasks. Claude Opus 5.5 topped the leaderboard at 34.3%, with GPT-6 models trailing by about 10 points.

2026-11-19 ~ 2026-11-21 · 2 related posts