AI models saturated the last evaluation, one post says
The post says the difficulty is being raised and asks whether models could reach 80% or higher by the end of 2026.
TLDR
A September 11 post says models had “saturated the last eval,” so the difficulty is being raised. It asks how soon they’ll reach 80% or higher on the harder evaluation, floating the end of 2026 as a possibility.
Combined views
—
1 Source, first seen ago
