Leak: Anthropic's unreleased Model 2 beats Mythos 5 on CoBench

Anthropic's unreleased Model 2 reportedly scores 62.8% on CoBench v2, beating Mythos 5 by 12.5 points and sitting just 22.2 points below the level Anthropic estimates would fully replace its researchers.

2026-08-15 ~ 2026-08-15 · 2 related posts

• Leak: Anthropic's Internal 'Model 2' Beats Mythos 5 by 12.5 Points on CoBench, Approaching Researcher-Level — daniel_mac8 · 2026-08-15
• Anthropic's unreleased Model 2 scores 62.8% on CoBench v2, closing gap on researchers — ChrisGPT · 2026-08-15