Reaction
Sonnet 5.5 could take three top spots across individual evaluations
A user contrasts Sonnet 5.5 with Opus 5.5, citing four top spots for Astra and one for 6.1 Sol.
TLDR
A user says Sonnet 5.5, unlike Opus 5.5, can take three top spots in individual evaluations, compared with four for Astra and one for 6.1 Sol. They also say Anthropic’s reinforcement learning is evolving quickly.
Combined views
793
1 Source, first seen ago
